Abstract
Discovering genes with similar functions across diverse biomedical contexts poses a significant challenge in gene representation learning due to data heterogeneity. In this study, we resolve this problem by introducing a novel model called Multimodal Similarity Learning Graph Neural Network, which combines Multimodal Machine Learning and Deep Graph Neural Networks to learn gene representations from single-cell sequencing and spatial transcriptomic data. Leveraging 82 training datasets from 10 tissues, three sequencing techniques, and three species, we create informative graph structures for model training and gene representations generation, while incorporating regularization with weighted similarity learning and contrastive learning to learn cross-data gene-gene relationships. This novel design ensures that we can offer gene representations containing functional similarity across different contexts in a joint space. Comprehensive benchmarking analysis shows our model's capacity to effectively capture gene function similarity across multiple modalities, outperforming state-of-the-art methods in gene representation learning by up to 97.5%. Moreover, we employ bioinformatics tools in conjunction with gene representations to uncover pathway enrichment, regulation causal networks, and functions of disease-associated or dosage-sensitive genes. Therefore, our model efficiently produces unified gene representations for the analysis of gene functions, tissue functions, diseases, and species evolution.
Consistency Trajectory Models: Learning Probability Flow ODE Trajectory of Diffusion
Abstract
Consistency Models (CM) (Song et al., 2023) accelerate score-based diffusion model sampling at the cost of sample quality but lack a natural way to trade-off quality for speed. To address this limitation, we propose Consistency Trajectory Model (CTM), a generalization encompassing CM and score-based models as special cases. CTM trains a single neural network that can -- in a single forward pass -- output scores (i.e., gradients of log-density) and enables unrestricted traversal between any initial and final time along the Probability Flow Ordinary Differential Equation (ODE) in a diffusion process. CTM enables the efficient combination of adversarial training and denoising score matching loss to enhance performance and achieves new state-of-the-art FIDs for single-step diffusion model sampling on CIFAR-10 (FID 1.73) and ImageNet at 64X64 resolution (FID 2.06). CTM also enables a new family of sampling schemes, both deterministic and stochastic, involving long jumps along the ODE solution trajectories. It consistently improves sample quality as computational budgets increase, avoiding the degradation seen in CM. Furthermore, CTM's access to the score accommodates all diffusion model inference techniques, including exact likelihood computation.
Expert enhanced dynamic time warping based anomaly detection
Authors: Matej Kloska, Gabriela Grmanova, Viera Rozinajova
Abstract
Dynamic time warping (DTW) is a well-known algorithm for time series elastic dissimilarity measure. Its ability to deal with non-linear time distortions makes it helpful in variety of data mining tasks. Such a task is also anomaly detection which attempts to reveal unexpected behaviour without false detection alarms. In this paper, we propose a novel anomaly detection method named Expert enhanced dynamic time warping anomaly detection (E-DTWA). It is based on DTW with additional enhancements involving human-in-the-loop concept. The main benefits of our approach comprise efficient detection, flexible retraining based on strong consideration of the expert's detection feedback while retaining low computational and space complexity.
A Comparison of Mesh-Free Differentiable Programming and Data-Driven Strategies for Optimal Control under PDE Constraints
Authors: Roussel Desmond Nzoyem, David A.W. Barton, Tom Deakin
Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Optimization and Control (math.OC)
Abstract
The field of Optimal Control under Partial Differential Equations (PDE) constraints is rapidly changing under the influence of Deep Learning and the accompanying automatic differentiation libraries. Novel techniques like Physics-Informed Neural Networks (PINNs) and Differentiable Programming (DP) are to be contrasted with established numerical schemes like Direct-Adjoint Looping (DAL). We present a comprehensive comparison of DAL, PINN, and DP using a general-purpose mesh-free differentiable PDE solver based on Radial Basis Functions. Under Laplace and Navier-Stokes equations, we found DP to be extremely effective as it produces the most accurate gradients; thriving even when DAL fails and PINNs struggle. Additionally, we provide a detailed benchmark highlighting the limited conditions under which any of those methods can be efficiently used. Our work provides a guide to Optimal Control practitioners and connects them further to the Deep Learning community.
Online Proactive Multi-Task Assignment with Resource Availability Anticipation
Abstract
With the emergence of services and online applications as taxi dispatching, crowdsourcing, package or food delivery, industrials and researchers are paying attention to the online multi-task assignment optimization field to quickly and efficiently met demands. In this context, this paper is interested in the multi-task assignment problem where multiple requests (e.g. tasks) arrive over time and must be dynamically matched to (mobile) agents. This optimization problem is known to be NP-hard. In order to treat this problem with a proactive mindset, we propose to use a receding-horizon approach to determine which resources (e.g. taxis, mobile agents, drones, robots) would be available within this (possibly dynamic) receding-horizon to meet the current set of requests (i.e. tasks) as good as possible. Contrarily to several works in this domain, we have chosen to make no assumption concerning future locations of requests. To achieve fast optimized online solutions in terms of costs and amount of allocated tasks, we have designed a genetic algorithm based on a fitness function integrating the traveled distance and the age of the requests. We compared our proactive multi-task assignment with resource availability anticipation approach with a classical reactive approach. The results obtained in two benchmark problems, one synthetic and another based on real data, show that our resource availability anticipation method can achieve better results in terms of costs (e.g. traveled distance) and amount of allocated tasks than reactive approaches while decreasing resources idle time.
Event-Enhanced Multi-Modal Spiking Neural Network for Dynamic Obstacle Avoidance
Authors: Yang Wang, Bo Dong, Yuji Zhang, Yunduo Zhou, Haiyang Mei, Ziqi Wei, Xin Yang
Abstract
Autonomous obstacle avoidance is of vital importance for an intelligent agent such as a mobile robot to navigate in its environment. Existing state-of-the-art methods train a spiking neural network (SNN) with deep reinforcement learning (DRL) to achieve energy-efficient and fast inference speed in complex/unknown scenes. These methods typically assume that the environment is static while the obstacles in real-world scenes are often dynamic. The movement of obstacles increases the complexity of the environment and poses a great challenge to the existing methods. In this work, we approach robust dynamic obstacle avoidance twofold. First, we introduce the neuromorphic vision sensor (i.e., event camera) to provide motion cues complementary to the traditional Laser depth data for handling dynamic obstacles. Second, we develop an DRL-based event-enhanced multimodal spiking actor network (EEM-SAN) that extracts information from motion events data via unsupervised representation learning and fuses Laser and event camera data with learnable thresholding. Experiments demonstrate that our EEM-SAN outperforms state-of-the-art obstacle avoidance methods by a significant margin, especially for dynamic obstacle avoidance.
Linear Recurrent Units for Sequential Recommendation
Abstract
State-of-the-art sequential recommendation relies heavily on self-attention-based recommender models. Yet such models are computationally expensive and often too slow for real-time recommendation. Furthermore, the self-attention operation is performed at a sequence-level, thereby making low-cost incremental inference challenging. Inspired by recent advances in efficient language modeling, we propose linear recurrent units for sequential recommendation (LRURec). Similar to recurrent neural networks, LRURec offers rapid inference and can achieve incremental inference on sequential inputs. By decomposing the linear recurrence operation and designing recursive parallelization in our framework, LRURec provides the additional benefits of reduced model size and parallelizable training. Moreover, we optimize the architecture of LRURec by implementing a series of modifications to address the lack of non-linearity and improve training dynamics. To validate the effectiveness of our proposed LRURec, we conduct extensive experiments on multiple real-world datasets and compare its performance against state-of-the-art sequential recommenders. Experimental results demonstrate the effectiveness of LRURec, which consistently outperforms baselines by a significant margin. Results also highlight the efficiency of LRURec with our parallelized training paradigm and fast inference on long sequences, showing its potential to further enhance user experience in sequential recommendation.
ScaleNet: An Unsupervised Representation Learning Method for Limited Information
Authors: Huili Huang, M. Mahdi Roozbahani
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Although large-scale labeled data are essential for deep convolutional neural networks (ConvNets) to learn high-level semantic visual representations, it is time-consuming and impractical to collect and annotate large-scale datasets. A simple and efficient unsupervised representation learning method named ScaleNet based on multi-scale images is proposed in this study to enhance the performance of ConvNets when limited information is available. The input images are first resized to a smaller size and fed to the ConvNet to recognize the rotation degree. Next, the ConvNet learns the rotation-prediction task for the original size images based on the parameters transferred from the previous model. The CIFAR-10 and ImageNet datasets are examined on different architectures such as AlexNet and ResNet50 in this study. The current study demonstrates that specific image features, such as Harris corner information, play a critical role in the efficiency of the rotation-prediction task. The ScaleNet supersedes the RotNet by ~7% in the limited CIFAR-10 dataset. The transferred parameters from a ScaleNet model with limited data improve the ImageNet Classification task by about 6% compared to the RotNet model. This study shows the capability of the ScaleNet method to improve other cutting-edge models such as SimCLR by learning effective features for classification tasks.
Mixture of Quantized Experts (MoQE): Complementary Effect of Low-bit Quantization and Robustness
Authors: Young Jin Kim, Raffy Fahim, Hany Hassan Awadalla
Subjects: Machine Learning (cs.LG); Computation and Language (cs.CL)
Abstract
Large Mixture of Experts (MoE) models could achieve state-of-the-art quality on various language tasks, including machine translation task, thanks to the efficient model scaling capability with expert parallelism. However, it has brought a fundamental issue of larger memory consumption and increased memory bandwidth bottleneck at deployment time. In this paper, we propose Mixture of Quantized Experts (MoQE) which is a simple weight-only quantization method applying ultra low-bit down to 2-bit quantizations only to expert weights for mitigating the increased memory and latency issues of MoE models. We show that low-bit quantization together with the MoE architecture delivers a reliable model performance while reducing the memory size significantly even without any additional training in most cases. In particular, expert layers in MoE models are much more robust to the quantization than conventional feedforward networks (FFN) layers. In our comprehensive analysis, we show that MoE models with 2-bit expert weights can deliver better model performance than the dense model trained on the same dataset. As a result of low-bit quantization, we show the model size can be reduced by 79.6% of the original half precision floating point (fp16) MoE model. Combined with an optimized GPU runtime implementation, it also achieves 1.24X speed-up on A100 GPUs.
Machine learning assist nyc subway navigation safer and faster
Authors: Wencheng Bao, Shi Feng
Subjects: Social and Information Networks (cs.SI); Computers and Society (cs.CY); Machine Learning (cs.LG)
Abstract
Mainstream navigation software, like Google and Apple Maps, often lacks the ability to provide routes prioritizing safety. However, safety remains a paramount concern for many. Our aim is to strike a balance between safety and efficiency. To achieve this, we're devising an Integer Programming model that takes into account both the shortest path and the safest route. We will harness machine learning to derive safety coefficients, employing methodologies such as generalized linear models, linear regression, and recurrent neural networks. Our evaluation will be based on the Root Mean Square Error (RMSE) across various subway stations, helping us identify the most accurate model for safety coefficient estimation. Furthermore, we'll conduct a comprehensive review of different shortest-path algorithms, assessing them based on time complexity and real-world data to determine their appropriateness in merging both safety and time efficiency.
Feather: An Elegant Solution to Effective DNN Sparsification
Authors: Athanasios Glentis Georgoulakis, George Retsinas, Petros Maragos
Abstract
Neural Network pruning is an increasingly popular way for producing compact and efficient models, suitable for resource-limited environments, while preserving high performance. While the pruning can be performed using a multi-cycle training and fine-tuning process, the recent trend is to encompass the sparsification process during the standard course of training. To this end, we introduce Feather, an efficient sparse training module utilizing the powerful Straight-Through Estimator as its core, coupled with a new thresholding operator and a gradient scaling technique, enabling robust, out-of-the-box sparsification performance. Feather's effectiveness and adaptability is demonstrated using various architectures on the CIFAR dataset, while on ImageNet it achieves state-of-the-art Top-1 validation accuracy using the ResNet-50 architecture, surpassing existing methods, including more complex and computationally heavy ones, by a considerable margin. Code is publicly available at https://github.com/athglentis/feather .
Impact of geography on the importance of parameters in infectious disease models
Abstract
Agent-based models are widely used to predict infectious disease spread. For these predictions, one needs to understand how each input parameter affects the result. Here, some parameters may affect the sensitivities of others, requiring the analysis of higher order coefficients through e.g. Sobol sensitivity analysis. The geographical structures of real-world regions are distinct in that they are difficult to reduce to single parameter values, making a unified sensitivity analysis intractable. Yet analyzing the importance of geographical structure on the sensitivity of other input parameters is important because a strong effect would justify the use of models with real-world geographical representations, as opposed to stylized ones. Here we perform a grouped Sobol's sensitivity analysis on COVID-19 spread simulations across a set of three diverse real-world geographical representations. We study the differences in both results and the sensitivity of non-geographical parameters across these geographies. By comparing Sobol indices of parameters across geographies, we find evidence that infection rate could have more sensitivity in regions where the population is segregated, while parameters like recovery period of mild cases are more sensitive in regions with mixed populations. We also show how geographical structure affects parameter sensitivity changes over time.
Distributionally Safe Reinforcement Learning under Model Uncertainty: A Single-Level Approach by Differentiable Convex Programming
Authors: Alaa Eddine Chriat, Chuangchuang Sun
Subjects: Machine Learning (cs.LG); Robotics (cs.RO); Systems and Control (eess.SY)
Abstract
Safety assurance is uncompromisable for safety-critical environments with the presence of drastic model uncertainties (e.g., distributional shift), especially with humans in the loop. However, incorporating uncertainty in safe learning will naturally lead to a bi-level problem, where at the lower level the (worst-case) safety constraint is evaluated within the uncertainty ambiguity set. In this paper, we present a tractable distributionally safe reinforcement learning framework to enforce safety under a distributional shift measured by a Wasserstein metric. To improve the tractability, we first use duality theory to transform the lower-level optimization from infinite-dimensional probability space where distributional shift is measured, to a finite-dimensional parametric space. Moreover, by differentiable convex programming, the bi-level safe learning problem is further reduced to a single-level one with two sequential computationally efficient modules: a convex quadratic program to guarantee safety followed by a projected gradient ascent to simultaneously find the worst-case uncertainty. This end-to-end differentiable framework with safety constraints, to the best of our knowledge, is the first tractable single-level solution to address distributional safety. We test our approach on first and second-order systems with varying complexities and compare our results with the uncertainty-agnostic policies, where our approach demonstrates a significant improvement on safety guarantees.
Parameterized Model-checking of Discrete-Timed Networks and Symmetric-Broadcast Systems
Authors: Benjamin Aminof, Sasha Rubin, Francesco Spegni, Florian Zuleger
Abstract
We study the complexity of the model-checking problem for discrete-timed systems with arbitrarily many anonymous and identical contributors, with and without a distinguished "controller" process, communicating via synchronous rendezvous. Our work extends the seminal work on untimed systems by German and Sistla adding discrete-time clocks, thus allowing one to model more realistic protocols. For the case without a controller, we show that the systems can be efficiently simulated -- and vice versa -- by systems of untimed processes communicating via rendezvous and symmetric broadcast, which we call "RB-systems". Symmetric broadcast is a novel communication primitive that, like ordinary asymmetric broadcast, allows all processes to synchronize without distinction between sender/receiver processes. We show that the complexity of the parameterized model-checking problem for safety specifications is pspace-complete, and for liveness specifications it is decidable in exptime. The latter result required automata theory, rational linear programming, and geometric reasoning for solving certain reachability questions in a new variant of vector addition systems called "vector rendezvous systems". We believe such proof techniques are of independent interest and will be useful in solving related problems. For the case with a controller, we show that the parameterized model-checking problems for RB-systems and systems with asymmetric broadcast are inter-reducible. This implies that for discrete timed-networks with a controller the parameterized model-checking problem is undecidable for liveness specifications. Our work exploits the intimate connection between discrete-timed systems and systems of processes communicating via broadcast. This allows us to prove decidability results for liveness properties of parameterized timed-systems, as well as extend work from untimed systems to timed systems.
Abstract
Machine learning traditionally assumes that training and testing data are distributed independently and identically. However, in many real-world settings, the data distribution can shift over time, leading to poor generalization of trained models in future time periods. Our paper presents a novel prompting-based approach to temporal domain generalization that is parameter-efficient, time-efficient, and does not require access to the target domain data (i.e., unseen future time periods) during training. Our method adapts a target pre-trained model to temporal drift by learning global prompts, domain-specific prompts, and drift-aware prompts that capture underlying temporal dynamics. It is compatible across diverse tasks, such as classification, regression, and time series forecasting, and sets a new state-of-the-art benchmark in temporal domain generalization. The code repository will be publicly shared.
Computationally Efficient Chance Constrained Covariance Control with Output Feedback
Authors: Joshua Pilipovsky, Panagiotis Tsiotras
Subjects: Systems and Control (eess.SY); Optimization and Control (math.OC)
Abstract
This paper studies the problem of developing computationally efficient solutions for steering the distribution of the state of a stochastic, linear dynamical system between two boundary Gaussian distributions in the presence of chance-constraints on the state and control input. It is assumed that the state is only partially available through a measurement model corrupted with noise. The filtered state is reconstructed with a Kalman filter, the chance constraints are reformulated as difference of convex (DC) constraints, and the resulting covariance control problem is reformulated as a DC program, which is solved using successive convexification. The efficiency of the proposed method is illustrated on a double integrator example with varying time horizons, and is compared to other state-of-the-art chance constrained covariance control methods.
Ophiuchus: Scalable Modeling of Protein Structures through Hierarchical Coarse-graining SO(3)-Equivariant Autoencoders
Authors: Allan dos Santos Costa, Ilan Mitnikov, Mario Geiger, Manvitha Ponnapati, Tess Smidt, Joseph Jacobson
Abstract
Three-dimensional native states of natural proteins display recurring and hierarchical patterns. Yet, traditional graph-based modeling of protein structures is often limited to operate within a single fine-grained resolution, and lacks hourglass neural architectures to learn those high-level building blocks. We narrow this gap by introducing Ophiuchus, an SO(3)-equivariant coarse-graining model that efficiently operates on all heavy atoms of standard protein residues, while respecting their relevant symmetries. Our model departs from current approaches that employ graph modeling, instead focusing on local convolutional coarsening to model sequence-motif interactions in log-linear length complexity. We train Ophiuchus on contiguous fragments of PDB monomers, investigating its reconstruction capabilities across different compression rates. We examine the learned latent space and demonstrate its prompt usage in conformational interpolation, comparing interpolated trajectories to structure snapshots from the PDBFlex dataset. Finally, we leverage denoising diffusion probabilistic models (DDPM) to efficiently sample readily-decodable latent embeddings of diverse miniproteins. Our experiments demonstrate Ophiuchus to be a scalable basis for efficient protein modeling and generation.
Abstract
Conditional stochastic optimization has found applications in a wide range of machine learning tasks, such as invariant learning, AUPRC maximization, and meta-learning. As the demand for training models with large-scale distributed data grows in these applications, there is an increasing need for communication-efficient distributed optimization algorithms, such as federated learning algorithms. This paper considers the nonconvex conditional stochastic optimization in federated learning and proposes the first federated conditional stochastic optimization algorithm (FCSG) with a conditional stochastic gradient estimator and a momentum-based algorithm (FCSG-M). To match the lower bound complexity in the single-machine setting, we design an accelerated algorithm (Acc-FCSG-M) via the variance reduction to achieve the best sample and communication complexity. Compared with the existing optimization analysis for MAML in FL, federated conditional stochastic optimization considers the sample of tasks. Extensive experimental results on various tasks validate the efficiency of these algorithms.
SlowFormer: Universal Adversarial Patch for Attack on Compute and Energy Efficiency of Inference Efficient Vision Transformers
Abstract
Recently, there has been a lot of progress in reducing the computation of deep models at inference time. These methods can reduce both the computational needs and power usage of deep models. Some of these approaches adaptively scale the compute based on the input instance. We show that such models can be vulnerable to a universal adversarial patch attack, where the attacker optimizes for a patch that when pasted on any image, can increase the compute and power consumption of the model. We run experiments with three different efficient vision transformer methods showing that in some cases, the attacker can increase the computation to the maximum possible level by simply pasting a patch that occupies only 8\% of the image area. We also show that a standard adversarial training defense method can reduce some of the attack's success. We believe adaptive efficient methods will be necessary for the future to lower the power usage of deep models, so we hope our paper encourages the community to study the robustness of these methods and develop better defense methods for the proposed attack.
Exact and soft boundary conditions in Physics-Informed Neural Networks for the Variable Coefficient Poisson equation
Abstract
Boundary conditions (BCs) are a key component in every Physics-Informed Neural Network (PINN). By defining the solution to partial differential equations (PDEs) along domain boundaries, BCs constrain the underlying boundary value problem (BVP) that a PINN tries to approximate. Without them, unique PDE solutions may not exist and finding approximations with PINNs would be a challenging, if not impossible task. This study examines how soft loss-based and exact distance function-based BC imposition approaches differ when applied in PINNs. The well known variable coefficient Poisson equation serves as the target PDE for all PINN models trained in this work. Besides comparing BC imposition approaches, the goal of this work is to also provide resources on how to implement these PINNs in practice. To this end, Keras models with Tensorflow backend as well as a Python notebook with code examples and step-by-step explanations on how to build soft/exact BC PINNs are published alongside this review.
Convergence Analysis and Latency Minimization for Semi-Federated Learning in Massive IoT Networks
Authors: Jianyang Ren, Wanli Ni, Hui Tian, Gaofeng Nie
Subjects: Information Theory (cs.IT); Signal Processing (eess.SP)
Abstract
As the number of sensors becomes massive in Internet of Things (IoT) networks, the amount of data is humongous. To process data in real-time while protecting user privacy, federated learning (FL) has been regarded as an enabling technique to push edge intelligence into IoT networks with massive devices. However, FL latency increases dramatically due to the increase of the number of parameters in deep neural network and the limited computation and communication capabilities of IoT devices. To address this issue, we propose a semi-federated learning (SemiFL) paradigm in which network pruning and over-the-air computation are efficiently applied. To be specific, each small base station collects the raw data from its served sensors and trains its local pruned model. After that, the global aggregation of local gradients is achieved through over-the-air computation. We first analyze the performance of the proposed SemiFL by deriving its convergence upper bound. To reduce latency, a convergence-constrained SemiFL latency minimization problem is formulated. By decoupling the original problem into several sub-problems, iterative algorithms are designed to solve them efficiently. Finally, numerical simulations are conducted to verify the effectiveness of our proposed scheme in reducing latency and guaranteeing the identification accuracy.
NOLA: Networks as Linear Combination of Low Rank Random Basis
Abstract
Large Language Models (LLMs) have recently gained popularity due to their impressive few-shot performance across various downstream tasks. However, fine-tuning all parameters and storing a unique model for each downstream task or domain becomes impractical because of the massive size of checkpoints (e.g., 350GB in GPT-3). Current literature, such as LoRA, showcases the potential of low-rank modifications to the original weights of an LLM, enabling efficient adaptation and storage for task-specific models. These methods can reduce the number of parameters needed to fine-tune an LLM by several orders of magnitude. Yet, these methods face two primary limitations: 1) the parameter reduction is lower-bounded by the rank one decomposition, and 2) the extent of reduction is heavily influenced by both the model architecture and the chosen rank. For instance, in larger models, even a rank one decomposition might exceed the number of parameters truly needed for adaptation. In this paper, we introduce NOLA, which overcomes the rank one lower bound present in LoRA. It achieves this by re-parameterizing the low-rank matrices in LoRA using linear combinations of randomly generated matrices (basis) and optimizing the linear mixture coefficients only. This approach allows us to decouple the number of trainable parameters from both the choice of rank and the network architecture. We present adaptation results using GPT-2 and ViT in natural language and computer vision tasks. NOLA performs as well as, or better than models with equivalent parameter counts. Furthermore, we demonstrate that we can halve the parameters in larger models compared to LoRA with rank one, without sacrificing performance.
Improving Drumming Robot Via Attention Transformer Network
Abstract
Robotic technology has been widely used in nowadays society, which has made great progress in various fields such as agriculture, manufacturing and entertainment. In this paper, we focus on the topic of drumming robots in entertainment. To this end, we introduce an improving drumming robot that can automatically complete music transcription based on the popular vision transformer network based on the attention mechanism. Equipped with the attention transformer network, our method can efficiently handle the sequential audio embedding input and model their global long-range dependencies. Massive experimental results demonstrate that the improving algorithm can help the drumming robot promote drum classification performance, which can also help the robot to enjoy a variety of smart applications and services.
ReForm-Eval: Evaluating Large Vision Language Models via Unified Re-Formulation of Task-Oriented Benchmarks
Authors: Zejun Li, Ye Wang, Mengfei Du, Qingwen Liu, Binhao Wu, Jiwen Zhang, Chengxing Zhou, Zhihao Fan, Jie Fu, Jingjing Chen, Xuanjing Huang, Zhongyu Wei
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Recent years have witnessed remarkable progress in the development of large vision-language models (LVLMs). Benefiting from the strong language backbones and efficient cross-modal alignment strategies, LVLMs exhibit surprising capabilities to perceive visual signals and perform visually grounded reasoning. However, the capabilities of LVLMs have not been comprehensively and quantitatively evaluate. Most existing multi-modal benchmarks require task-oriented input-output formats, posing great challenges to automatically assess the free-form text output of LVLMs. To effectively leverage the annotations available in existing benchmarks and reduce the manual effort required for constructing new benchmarks, we propose to re-formulate existing benchmarks into unified LVLM-compatible formats. Through systematic data collection and reformulation, we present the ReForm-Eval benchmark, offering substantial data for evaluating various capabilities of LVLMs. Based on ReForm-Eval, we conduct extensive experiments, thoroughly analyze the strengths and weaknesses of existing LVLMs, and identify the underlying factors. Our benchmark and evaluation framework will be open-sourced as a cornerstone for advancing the development of LVLMs.
AdaMerging: Adaptive Model Merging for Multi-Task Learning
Authors: Enneng Yang, Zhenyi Wang, Li Shen, Shiwei Liu, Guibing Guo, Xingwei Wang, Dacheng Tao
Subjects: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
Abstract
Multi-task learning (MTL) aims to empower a model to tackle multiple tasks simultaneously. A recent development known as task arithmetic has revealed that several models, each fine-tuned for distinct tasks, can be directly merged into a single model to execute MTL without necessitating a retraining process using the initial training data. Nevertheless, this direct addition of models often leads to a significant deterioration in the overall performance of the merged model. This decline occurs due to potential conflicts and intricate correlations among the multiple tasks. Consequently, the challenge emerges of how to merge pre-trained models more effectively without using their original training data. This paper introduces an innovative technique called Adaptive Model Merging (AdaMerging). This approach aims to autonomously learn the coefficients for model merging, either in a task-wise or layer-wise manner, without relying on the original training data. Specifically, our AdaMerging method operates as an automatic, unsupervised task arithmetic scheme. It leverages entropy minimization on unlabeled test samples from the multi-task setup as a surrogate objective function to iteratively refine the merging coefficients of the multiple models. Our experimental findings across eight tasks demonstrate the efficacy of the AdaMerging scheme we put forth. Compared to the current state-of-the-art task arithmetic merging scheme, AdaMerging showcases a remarkable 11\% improvement in performance. Notably, AdaMerging also exhibits superior generalization capabilities when applied to unseen downstream tasks. Furthermore, it displays a significantly enhanced robustness to data distribution shifts that may occur during the testing phase.
ViT-ReciproCAM: Gradient and Attention-Free Visual Explanations for Vision Transformer
Authors: Seok-Yong Byun, Wonju Lee
Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Abstract
This paper presents a novel approach to address the challenges of understanding the prediction process and debugging prediction errors in Vision Transformers (ViT), which have demonstrated superior performance in various computer vision tasks such as image classification and object detection. While several visual explainability techniques, such as CAM, Grad-CAM, Score-CAM, and Recipro-CAM, have been extensively researched for Convolutional Neural Networks (CNNs), limited research has been conducted on ViT. Current state-of-the-art solutions for ViT rely on class agnostic Attention-Rollout and Relevance techniques. In this work, we propose a new gradient-free visual explanation method for ViT, called ViT-ReciproCAM, which does not require attention matrix and gradient information. ViT-ReciproCAM utilizes token masking and generated new layer outputs from the target layer's input to exploit the correlation between activated tokens and network predictions for target classes. Our proposed method outperforms the state-of-the-art Relevance method in the Average Drop-Coherence-Complexity (ADCC) metric by $4.58\%$ to $5.80\%$ and generates more localized saliency maps. Our experiments demonstrate the effectiveness of ViT-ReciproCAM and showcase its potential for understanding and debugging ViT models. Our proposed method provides an efficient and easy-to-implement alternative for generating visual explanations, without requiring attention and gradient information, which can be beneficial for various applications in the field of computer vision.
Hybrid Inception Architecture with Residual Connection: Fine-tuned Inception-ResNet Deep Learning Model for Lung Inflammation Diagnosis from Chest Radiographs
Abstract
Diagnosing lung inflammation, particularly pneumonia, is of paramount importance for effectively treating and managing the disease. Pneumonia is a common respiratory infection caused by bacteria, viruses, or fungi and can indiscriminately affect people of all ages. As highlighted by the World Health Organization (WHO), this prevalent disease tragically accounts for a substantial 15% of global mortality in children under five years of age. This article presents a comparative study of the Inception-ResNet deep learning model's performance in diagnosing pneumonia from chest radiographs. The study leverages Mendeleys chest X-ray images dataset, which contains 5856 2D images, including both Viral and Bacterial Pneumonia X-ray images. The Inception-ResNet model is compared with seven other state-of-the-art convolutional neural networks (CNNs), and the experimental results demonstrate the Inception-ResNet model's superiority in extracting essential features and saving computation runtime. Furthermore, we examine the impact of transfer learning with fine-tuning in improving the performance of deep convolutional models. This study provides valuable insights into using deep learning models for pneumonia diagnosis and highlights the potential of the Inception-ResNet model in this field. In classification accuracy, Inception-ResNet-V2 showed superior performance compared to other models, including ResNet152V2, MobileNet-V3 (Large and Small), EfficientNetV2 (Large and Small), InceptionV3, and NASNet-Mobile, with substantial margins. It outperformed them by 2.6%, 6.5%, 7.1%, 13%, 16.1%, 3.9%, and 1.6%, respectively, demonstrating its significant advantage in accurate classification.
Multi-Agent Reinforcement Learning for Power Grid Topology Optimization
Authors: Erica van der Sar, Alessandro Zocca, Sandjai Bhulai
Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Systems and Control (eess.SY); Machine Learning (stat.ML)
Abstract
Recent challenges in operating power networks arise from increasing energy demands and unpredictable renewable sources like wind and solar. While reinforcement learning (RL) shows promise in managing these networks, through topological actions like bus and line switching, efficiently handling large action spaces as networks grow is crucial. This paper presents a hierarchical multi-agent reinforcement learning (MARL) framework tailored for these expansive action spaces, leveraging the power grid's inherent hierarchical nature. Experimental results indicate the MARL framework's competitive performance with single-agent RL methods. We also compare different RL algorithms for lower-level agents alongside different policies for higher-order agents.
Top-k contrast order-preserving pattern mining for time series classification
Authors: Youxi Wu, Yufei Meng, Yan Li, Lei Guo, Xingquan Zhu, Philippe Fournier-Viger, Xindong Wu
Abstract
Recently, order-preserving pattern (OPP) mining, a new sequential pattern mining method, has been proposed to mine frequent relative orders in a time series. Although frequent relative orders can be used as features to classify a time series, the mined patterns do not reflect the differences between two classes of time series well. To effectively discover the differences between time series, this paper addresses the top-k contrast OPP (COPP) mining and proposes a COPP-Miner algorithm to discover the top-k contrast patterns as features for time series classification, avoiding the problem of improper parameter setting. COPP-Miner is composed of three parts: extreme point extraction to reduce the length of the original time series, forward mining, and reverse mining to discover COPPs. Forward mining contains three steps: group pattern fusion strategy to generate candidate patterns, the support rate calculation method to efficiently calculate the support of a pattern, and two pruning strategies to further prune candidate patterns. Reverse mining uses one pruning strategy to prune candidate patterns and consists of applying the same process as forward mining. Experimental results validate the efficiency of the proposed algorithm and show that top-k COPPs can be used as features to obtain a better classification performance.
Multi-rules mining algorithm for combinatorially exploded decision trees with modified Aitchison-Aitken function-based Bayesian optimization
Abstract
Decision trees offer the benefit of easy interpretation because they allow the classification of input data based on if--then rules. However, as decision trees are constructed by an algorithm that achieves clear classification with minimum necessary rules, the trees possess the drawback of extracting only minimum rules, even when various latent rules exist in data. Approaches that construct multiple trees using randomly selected feature subsets do exist. However, the number of trees that can be constructed remains at the same scale because the number of feature subsets is a combinatorial explosion. Additionally, when multiple trees are constructed, numerous rules are generated, of which several are untrustworthy and/or highly similar. Therefore, we propose "MAABO-MT" and "GS-MRM" algorithms that strategically construct trees with high estimation performance among all possible trees with small computational complexity and extract only reliable and non-similar rules, respectively. Experiments are conducted using several open datasets to analyze the effectiveness of the proposed method. The results confirm that MAABO-MT can discover reliable rules at a lower computational cost than other methods that rely on randomness. Furthermore, the proposed method is confirmed to provide deeper insights than single decision trees commonly used in previous studies. Therefore, MAABO-MT and GS-MRM can efficiently extract rules from combinatorially exploded decision trees.
Foundation Reinforcement Learning: towards Embodied Generalist Agents with Foundation Prior Assistance
Authors: Weirui Ye, Yunsheng Zhang, Mengchen Wang, Shengjie Wang, Xianfan Gu, Pieter Abbeel, Yang Gao
Abstract
Recently, people have shown that large-scale pre-training from internet-scale data is the key to building generalist models, as witnessed in NLP. To build embodied generalist agents, we and many other researchers hypothesize that such foundation prior is also an indispensable component. However, it is unclear what is the proper concrete form to represent those embodied foundation priors and how they should be used in the downstream task. In this paper, we propose an intuitive and effective set of embodied priors that consist of foundation policy, value, and success reward. The proposed priors are based on the goal-conditioned MDP. To verify their effectiveness, we instantiate an actor-critic method assisted by the priors, called Foundation Actor-Critic (FAC). We name our framework as Foundation Reinforcement Learning (FRL), since it completely relies on embodied foundation priors to explore, learn and reinforce. The benefits of FRL are threefold. (1) Sample efficient. With foundation priors, FAC learns significantly faster than traditional RL. Our evaluation on the Meta-World has proved that FAC can achieve 100% success rates for 7/8 tasks under less than 200k frames, which outperforms the baseline method with careful manual-designed rewards under 1M frames. (2) Robust to noisy priors. Our method tolerates the unavoidable noise in embodied foundation models. We show that FAC works well even under heavy noise or quantization errors. (3) Minimal human intervention: FAC completely learns from the foundation priors, without the need of human-specified dense reward, or providing teleoperated demos. Thus, FAC can be easily scaled up. We believe our FRL framework could enable the future robot to autonomously explore and learn without human intervention in the physical world. In summary, our proposed FRL is a novel and powerful learning paradigm, towards achieving embodied generalist agents.
Deformation-Invariant Neural Network and Its Applications in Distorted Image Restoration and Analysis
Authors: Han Zhang, Qiguang Chen, Lok Ming Lui
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Image and Video Processing (eess.IV)
Abstract
Images degraded by geometric distortions pose a significant challenge to imaging and computer vision tasks such as object recognition. Deep learning-based imaging models usually fail to give accurate performance for geometrically distorted images. In this paper, we propose the deformation-invariant neural network (DINN), a framework to address the problem of imaging tasks for geometrically distorted images. The DINN outputs consistent latent features for images that are geometrically distorted but represent the same underlying object or scene. The idea of DINN is to incorporate a simple component, called the quasiconformal transformer network (QCTN), into other existing deep networks for imaging tasks. The QCTN is a deep neural network that outputs a quasiconformal map, which can be used to transform a geometrically distorted image into an improved version that is closer to the distribution of natural or good images. It first outputs a Beltrami coefficient, which measures the quasiconformality of the output deformation map. By controlling the Beltrami coefficient, the local geometric distortion under the quasiconformal mapping can be controlled. The QCTN is lightweight and simple, which can be readily integrated into other existing deep neural networks to enhance their performance. Leveraging our framework, we have developed an image classification network that achieves accurate classification of distorted images. Our proposed framework has been applied to restore geometrically distorted images by atmospheric turbulence and water turbulence. DINN outperforms existing GAN-based restoration methods under these scenarios, demonstrating the effectiveness of the proposed framework. Additionally, we apply our proposed framework to the 1-1 verification of human face images under atmospheric turbulence and achieve satisfactory performance, further demonstrating the efficacy of our approach.
Blend: A Unified Data Discovery System
Authors: Mahdi Esmailoghli, Christoph Schnell, Renée J. Miller, Ziawasch Abedjan
Abstract
Data discovery is an iterative and incremental process that necessitates the execution of multiple data discovery queries to identify the desired tables from large and diverse data lakes. Current methodologies concentrate on single discovery tasks such as join, correlation, or union discovery. However, in practice, a series of these approaches and their corresponding index structures are necessary to enable the user to discover the desired tables. This paper presents BLEND, a comprehensive data discovery system that empowers users to develop ad-hoc discovery tasks without the need to develop new algorithms or build a new index structure. To achieve this goal, we introduce a general index structure capable of addressing multiple discovery queries. We develop a set of lower-level operators that serve as the fundamental building blocks for more complex and sophisticated user tasks. These operators are highly efficient and enable end-to-end efficiency. To enhance the execution of the discovery pipeline, we rewrite the search queries into optimized SQL statements to push the data operators down to the database. We demonstrate that our holistic system is able to achieve comparable effectiveness and runtime efficiency to the individual state-of-the-art approaches specifically designed for a single task.
MedPrompt: Cross-Modal Prompting for Multi-Task Medical Image Translation
Authors: Xuhang Chen, Chi-Man Pun, Shuqiang Wang
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Cross-modal medical image translation is an essential task for synthesizing missing modality data for clinical diagnosis. However, current learning-based techniques have limitations in capturing cross-modal and global features, restricting their suitability to specific pairs of modalities. This lack of versatility undermines their practical usefulness, particularly considering that the missing modality may vary for different cases. In this study, we present MedPrompt, a multi-task framework that efficiently translates different modalities. Specifically, we propose the Self-adaptive Prompt Block, which dynamically guides the translation network towards distinct modalities. Within this framework, we introduce the Prompt Extraction Block and the Prompt Fusion Block to efficiently encode the cross-modal prompt. To enhance the extraction of global features across diverse modalities, we incorporate the Transformer model. Extensive experimental results involving five datasets and four pairs of modalities demonstrate that our proposed model achieves state-of-the-art visual quality and exhibits excellent generalization capability.
Land-cover change detection using paired OpenStreetMap data and optical high-resolution imagery via object-guided Transformer
Abstract
Optical high-resolution imagery and OpenStreetMap (OSM) data are two important data sources for land-cover change detection. Previous studies in these two data sources focus on utilizing the information in OSM data to aid the change detection on multi-temporal optical high-resolution images. This paper pioneers the direct detection of land-cover changes utilizing paired OSM data and optical imagery, thereby broadening the horizons of change detection tasks to encompass more dynamic earth observations. To this end, we propose an object-guided Transformer (ObjFormer) architecture by naturally combining the prevalent object-based image analysis (OBIA) technique with the advanced vision Transformer architecture. The introduction of OBIA can significantly reduce the computational overhead and memory burden in the self-attention module. Specifically, the proposed ObjFormer has a hierarchical pseudo-siamese encoder consisting of object-guided self-attention modules that extract representative features of different levels from OSM data and optical images; a decoder consisting of object-guided cross-attention modules can progressively recover the land-cover changes from the extracted heterogeneous features. In addition to the basic supervised binary change detection task, this paper raises a new semi-supervised semantic change detection task that does not require any manually annotated land-cover labels of optical images to train semantic change detectors. Two lightweight semantic decoders are added to ObjFormer to accomplish this task efficiently. A converse cross-entropy loss is designed to fully utilize the negative samples, thereby contributing to the great performance improvement in this task. The first large-scale benchmark dataset containing 1,287 map-image pairs (1024$\times$ 1024 pixels for each sample) covering 40 regions on six continents ...(see the manuscript for the full abstract)
GETAvatar: Generative Textured Meshes for Animatable Human Avatars
Abstract
We study the problem of 3D-aware full-body human generation, aiming at creating animatable human avatars with high-quality textures and geometries. Generally, two challenges remain in this field: i) existing methods struggle to generate geometries with rich realistic details such as the wrinkles of garments; ii) they typically utilize volumetric radiance fields and neural renderers in the synthesis process, making high-resolution rendering non-trivial. To overcome these problems, we propose GETAvatar, a Generative model that directly generates Explicit Textured 3D meshes for animatable human Avatar, with photo-realistic appearance and fine geometric details. Specifically, we first design an articulated 3D human representation with explicit surface modeling, and enrich the generated humans with realistic surface details by learning from the 2D normal maps of 3D scan data. Second, with the explicit mesh representation, we can use a rasterization-based renderer to perform surface rendering, allowing us to achieve high-resolution image generation efficiently. Extensive experiments demonstrate that GETAvatar achieves state-of-the-art performance on 3D-aware human generation both in appearance and geometry quality. Notably, GETAvatar can generate images at 512x512 resolution with 17FPS and 1024x1024 resolution with 14FPS, improving upon previous methods by 2x. Our code and models will be available.
Multi-resolution HuBERT: Multi-resolution Speech Self-Supervised Learning with Masked Unit Prediction
Authors: Jiatong Shi, Hirofumi Inaguma, Xutai Ma, Ilia Kulikov, Anna Sun
Subjects: Sound (cs.SD); Audio and Speech Processing (eess.AS)
Abstract
Existing Self-Supervised Learning (SSL) models for speech typically process speech signals at a fixed resolution of 20 milliseconds. This approach overlooks the varying informational content present at different resolutions in speech signals. In contrast, this paper aims to incorporate multi-resolution information into speech self-supervised representation learning. We introduce a SSL model that leverages a hierarchical Transformer architecture, complemented by HuBERT-style masked prediction objectives, to process speech at multiple resolutions. Experimental results indicate that the proposed model not only achieves more efficient inference but also exhibits superior or comparable performance to the original HuBERT model over various tasks. Specifically, significant performance improvements over the original HuBERT have been observed in fine-tuning experiments on the LibriSpeech speech recognition benchmark as well as in evaluations using the Speech Universal PERformance Benchmark (SUPERB) and Multilingual SUPERB (ML-SUPERB).
Leveraging Temporal Graph Networks Using Module Decoupling
Abstract
Modern approaches for learning on dynamic graphs have adopted the use of batches instead of applying updates one by one. The use of batches allows these techniques to become helpful in streaming scenarios where updates to graphs are received at extreme speeds. Using batches, however, forces the models to update infrequently, which results in the degradation of their performance. In this work, we suggest a decoupling strategy that enables the models to update frequently while using batches. By decoupling the core modules of temporal graph networks and implementing them using a minimal number of learnable parameters, we have developed the Lightweight Decoupled Temporal Graph Network (LDTGN), an exceptionally efficient model for learning on dynamic graphs. LDTG was validated on various dynamic graph benchmarks, providing comparable or state-of-the-art results with significantly higher throughput than previous art. Notably, our method outperforms previous approaches by more than 20\% on benchmarks that require rapid model update rates, such as USLegis or UNTrade. The code to reproduce our experiments is available at \href{https://orfeld415.github.io/module-decoupling}{this http url}.
Optimal Collaborative Transportation for Under-Capacitated Vehicle Routing Problems using Aerial Drone Swarms
Abstract
Swarms of aerial drones have recently been considered for last-mile deliveries in urban logistics or automated construction. At the same time, collaborative transportation of payloads by multiple drones is another important area of recent research. However, efficient coordination algorithms for collaborative transportation of many payloads by many drones remain to be considered. In this work, we formulate the collaborative transportation of payloads by a swarm of drones as a novel, under-capacitated generalization of vehicle routing problems (VRP), which may also be of separate interest. In contrast to standard VRP and capacitated VRP, we must additionally consider waiting times for payloads lifted cooperatively by multiple drones, and the corresponding coordination. Algorithmically, we provide a solution encoding that avoids deadlocks and formulate an appropriate alternating minimization scheme to solve the problem. On the hardware side, we integrate our algorithms with collision avoidance and drone controllers. The approach and the impact of the system integration are successfully verified empirically, both on a swarm of real nano-quadcopters and for large swarms in simulation. Overall, we provide a framework for collaborative transportation with aerial drone swarms, that uses only as many drones as necessary for the transportation of any single payload.
Comparative Analysis of Imbalanced Malware Byteplot Image Classification using Transfer Learning
Abstract
Cybersecurity is a major concern due to the increasing reliance on technology and interconnected systems. Malware detectors help mitigate cyber-attacks by comparing malware signatures. Machine learning can improve these detectors by automating feature extraction, identifying patterns, and enhancing dynamic analysis. In this paper, the performance of six multiclass classification models is compared on the Malimg dataset, Blended dataset, and Malevis dataset to gain insights into the effect of class imbalance on model performance and convergence. It is observed that the more the class imbalance less the number of epochs required for convergence and a high variance across the performance of different models. Moreover, it is also observed that for malware detectors ResNet50, EfficientNetB0, and DenseNet169 can handle imbalanced and balanced data well. A maximum precision of 97% is obtained for the imbalanced dataset, a maximum precision of 95% is obtained on the intermediate imbalance dataset, and a maximum precision of 95% is obtained for the perfectly balanced dataset.
Likelihood-Based Methods Improve Parameter Estimation in Opinion Dynamics Models
Authors: Jacopo Lenti, Gianmarco De Francisci Morales, Corrado Monti
Subjects: Social and Information Networks (cs.SI); Computers and Society (cs.CY)
Abstract
We show that a maximum likelihood approach for parameter estimation in agent-based models (ABMs) of opinion dynamics outperforms the typical simulation-based approach. Simulation-based approaches simulate the model repeatedly in search of a set of parameters that generates data similar enough to the observed one. In contrast, likelihood-based approaches derive a likelihood function that connects the unknown parameters to the observed data in a statistically principled way. We compare these two approaches on the well-known bounded-confidence model of opinion dynamics. We do so on three realistic scenarios of increasing complexity depending on data availability: (i) fully observed opinions and interactions, (ii) partially observed interactions, (iii) observed interactions with noisy proxies of the opinions. We highlight how identifying observed and latent variables is fundamental for connecting the model to the data. To realize the likelihood-based approach, we first cast the model into a probabilistic generative guise that supports a proper data likelihood. Then, we describe the three scenarios via probabilistic graphical models and show the nuances that go into translating the model. Finally, we implement the resulting probabilistic models in an automatic differentiation framework (PyTorch). This step enables easy and efficient maximum likelihood estimation via gradient descent. Our experimental results show that the maximum likelihood estimates are up to 4x more accurate and require up to 200x less computational time.
Spike Accumulation Forwarding for Effective Training of Spiking Neural Networks
Abstract
In this article, we propose a new paradigm for training spiking neural networks (SNNs), spike accumulation forwarding (SAF). It is known that SNNs are energy-efficient but difficult to train. Consequently, many researchers have proposed various methods to solve this problem, among which online training through time (OTTT) is a method that allows inferring at each time step while suppressing the memory cost. However, to compute efficiently on GPUs, OTTT requires operations with spike trains and weighted summation of spike trains during forwarding. In addition, OTTT has shown a relationship with the Spike Representation, an alternative training method, though theoretical agreement with Spike Representation has yet to be proven. Our proposed method can solve these problems; namely, SAF can halve the number of operations during the forward process, and it can be theoretically proven that SAF is consistent with the Spike Representation and OTTT, respectively. Furthermore, we confirmed the above contents through experiments and showed that it is possible to reduce memory and training time while maintaining accuracy.
Dynamic Shuffle: An Efficient Channel Mixture Method
Abstract
The redundancy of Convolutional neural networks not only depends on weights but also depends on inputs. Shuffling is an efficient operation for mixing channel information but the shuffle order is usually pre-defined. To reduce the data-dependent redundancy, we devise a dynamic shuffle module to generate data-dependent permutation matrices for shuffling. Since the dimension of permutation matrix is proportional to the square of the number of input channels, to make the generation process efficiently, we divide the channels into groups and generate two shared small permutation matrices for each group, and utilize Kronecker product and cross group shuffle to obtain the final permutation matrices. To make the generation process learnable, based on theoretical analysis, softmax, orthogonal regularization, and binarization are employed to asymptotically approximate the permutation matrix. Dynamic shuffle adaptively mixes channel information with negligible extra computation and memory occupancy. Experiment results on image classification benchmark datasets CIFAR-10, CIFAR-100, Tiny ImageNet and ImageNet have shown that our method significantly increases ShuffleNets' performance. Adding dynamic generated matrix with learnable static matrix, we further propose static-dynamic-shuffle and show that it can serve as a lightweight replacement of ordinary pointwise convolution.
R-LGP: A Reachability-guided Logic-geometric Programming Framework for Optimal Task and Motion Planning on Mobile Manipulators
Authors: Kim Tien Ly, Valeriy Semenov, Mattia Risiglione, Wolfgang Merkt, Ioannis Havoutis
Abstract
This paper presents an optimization-based solution to task and motion planning (TAMP) on mobile manipulators. Logic-geometric programming (LGP) has shown promising capabilities for optimally dealing with hybrid TAMP problems that involve abstract and geometric constraints. However, LGP does not scale well to high-dimensional systems (e.g. mobile manipulators) and can suffer from obstacle avoidance issues. In this work, we extend LGP with a sampling-based reachability graph to enable solving optimal TAMP on high-DoF mobile manipulators. The proposed reachability graph can incorporate environmental information (obstacles) to provide the planner with sufficient geometric constraints. This reachability-aware heuristic efficiently prunes infeasible sequences of actions in the continuous domain, hence, it reduces replanning by securing feasibility at the final full trajectory optimization. Our framework proves to be time-efficient in computing optimal and collision-free solutions, while outperforming the current state of the art on metrics of success rate, planning time, path length and number of steps. We validate our framework on the physical Toyota HSR robot and report comparisons on a series of mobile manipulation tasks of increasing difficulty.
Everest: GPU-Accelerated System For Mining Temporal Motifs
Abstract
Temporal motif mining is the task of finding the occurrences of subgraph patterns within a large input temporal graph that obey the specified structural and temporal constraints. Despite its utility in several critical application domains that demand high performance (e.g., detecting fraud in financial transaction graphs), the performance of existing software is limited on commercial hardware platforms, in that it runs for tens of hours. This paper presents Everest - a system that efficiently maps the workload of mining (supports both enumeration and counting) temporal motifs to the highly parallel GPU architecture. In particular, using an input temporal graph and a more expressive user-defined temporal motif query definition compared to prior works, Everest generates an execution plan and runtime primitives that optimize the workload execution by exploiting the high compute throughput of a GPU. Everest generates motif-specific mining code to reduce long-latency memory accesses and frequent thread divergence operations. Everest incorporates novel low-cost runtime mechanisms to enable load balancing to improve GPU hardware utilization. To support large graphs that do not fit on GPU memory, Everest also supports multi-GPU execution by intelligently partitioning the edge list that prevents inter-GPU communication. Everest hides the implementation complexity of presented optimizations away from the targeted system user for better usability. Our evaluation shows that, using proposed optimizations, Everest improves the performance of a baseline GPU implementation by 19x, on average.
Delving into CLIP latent space for Video Anomaly Recognition
Abstract
We tackle the complex problem of detecting and recognising anomalies in surveillance videos at the frame level, utilising only video-level supervision. We introduce the novel method AnomalyCLIP, the first to combine Large Language and Vision (LLV) models, such as CLIP, with multiple instance learning for joint video anomaly detection and classification. Our approach specifically involves manipulating the latent CLIP feature space to identify the normal event subspace, which in turn allows us to effectively learn text-driven directions for abnormal events. When anomalous frames are projected onto these directions, they exhibit a large feature magnitude if they belong to a particular class. We also introduce a computationally efficient Transformer architecture to model short- and long-term temporal dependencies between frames, ultimately producing the final anomaly score and class prediction probabilities. We compare AnomalyCLIP against state-of-the-art methods considering three major anomaly detection benchmarks, i.e. ShanghaiTech, UCF-Crime, and XD-Violence, and empirically show that it outperforms baselines in recognising video anomalies.
A novel asymmetrical autoencoder with a sparsifying discrete cosine Stockwell transform layer for gearbox sensor data compression
Abstract
The lack of an efficient compression model remains a challenge for the wireless transmission of gearbox data in non-contact gear fault diagnosis problems. In this paper, we present a signal-adaptive asymmetrical autoencoder with a transform domain layer to compress sensor signals. First, a new discrete cosine Stockwell transform (DCST) layer is introduced to replace linear layers in a multi-layer autoencoder. A trainable filter is implemented in the DCST domain by utilizing the multiplication property of the convolution. A trainable hard-thresholding layer is applied to reduce redundant data in the DCST layer to make the feature map sparse. In comparison to the linear layer, the DCST layer reduces the number of trainable parameters and improves the accuracy of data reconstruction. Second, training the autoencoder with a sparsifying DCST layer only requires a small number of datasets. The proposed method is superior to other autoencoder-based methods on the University of Connecticut (UoC) and Southeast University (SEU) gearbox datasets, as the average quality score is improved by 2.00% at the lowest and 32.35% at the highest with a limited number of training samples
Approximating Robot Configuration Spaces with few Convex Sets using Clique Covers of Visibility Graphs
Authors: Peter Werner, Alexandre Amice, Tobia Marcucci, Daniela Rus, Russ Tedrake
Abstract
Many computations in robotics can be dramatically accelerated if the robot configuration space is described as a collection of simple sets. For example, recently developed motion planners rely on a convex decomposition of the free space to design collision-free trajectories using fast convex optimization. In this work, we present an efficient method for approximately covering complex configuration spaces with a small number of polytopes. The approach constructs a visibility graph using sampling and generates a clique cover of this graph to find clusters of samples that have mutual line of sight. These clusters are then inflated into large, full-dimensional, polytopes. We evaluate our method on a variety of robotic systems and show that it consistently covers larger portions of free configuration space, with fewer polytopes, and in a fraction of the time compared to previous methods.
A Grammatical Compositional Model for Video Action Detection
Abstract
Analysis of human actions in videos demands understanding complex human dynamics, as well as the interaction between actors and context. However, these interaction relationships usually exhibit large intra-class variations from diverse human poses or object manipulations, and fine-grained inter-class differences between similar actions. Thus the performance of existing methods is severely limited. Motivated by the observation that interactive actions can be decomposed into actor dynamics and participating objects or humans, we propose to investigate the composite property of them. In this paper, we present a novel Grammatical Compositional Model (GCM) for action detection based on typical And-Or graphs. Our model exploits the intrinsic structures and latent relationships of actions in a hierarchical manner to harness both the compositionality of grammar models and the capability of expressing rich features of DNNs. The proposed model can be readily embodied into a neural network module for efficient optimization in an end-to-end manner. Extensive experiments are conducted on the AVA dataset and the Something-Else task to demonstrate the superiority of our model, meanwhile the interpretability is enhanced through an inference parsing procedure.
CoLiDE: Concomitant Linear DAG Estimation
Authors: Seyed Saman Saboksayr, Gonzalo Mateos, Mariano Tepper
Abstract
We deal with the combinatorial problem of learning directed acyclic graph (DAG) structure from observational data adhering to a linear structural equation model (SEM). Leveraging advances in differentiable, nonconvex characterizations of acyclicity, recent efforts have advocated a continuous constrained optimization paradigm to efficiently explore the space of DAGs. Most existing methods employ lasso-type score functions to guide this search, which (i) require expensive penalty parameter retuning when the $\textit{unknown}$ SEM noise variances change across problem instances; and (ii) implicitly rely on limiting homoscedasticity assumptions. In this work, we propose a new convex score function for sparsity-aware learning of linear DAGs, which incorporates concomitant estimation of scale and thus effectively decouples the sparsity parameter from the exogenous noise levels. Regularization via a smooth, nonconvex acyclicity penalty term yields CoLiDE ($\textbf{Co}$ncomitant $\textbf{Li}$near $\textbf{D}$AG $\textbf{E}$stimation), a regression-based criterion amenable to efficient gradient computation and closed-form estimation of noise variances in heteroscedastic scenarios. Our algorithm outperforms state-of-the-art methods without incurring added complexity, especially when the DAGs are larger and the noise level profile is heterogeneous. We also find CoLiDE exhibits enhanced stability manifested via reduced standard deviations in several domain-specific metrics, underscoring the robustness of our novel linear DAG estimator.
Use Your INSTINCT: INSTruction optimization usIng Neural bandits Coupled with Transformers
Authors: Xiaoqiang Lin, Zhaoxuan Wu, Zhongxiang Dai, Wenyang Hu, Yao Shu, See-Kiong Ng, Patrick Jaillet, Bryan Kian Hsiang Low
Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
Abstract
Large language models (LLMs) have shown remarkable instruction-following capabilities and achieved impressive performances in various applications. However, the performances of LLMs depend heavily on the instructions given to them, which are typically manually tuned with substantial human efforts. Recent work has used the query-efficient Bayesian optimization (BO) algorithm to automatically optimize the instructions given to black-box LLMs. However, BO usually falls short when optimizing highly sophisticated (e.g., high-dimensional) objective functions, such as the functions mapping an instruction to the performance of an LLM. This is mainly due to the limited expressive power of the Gaussian process (GP) model which is used by BO as a surrogate to model the objective function. Meanwhile, it has been repeatedly shown that neural networks (NNs), especially pre-trained transformers, possess strong expressive power and can model highly complex functions. So, we adopt a neural bandit algorithm which replaces the GP in BO by an NN surrogate to optimize instructions for black-box LLMs. More importantly, the neural bandit algorithm allows us to naturally couple the NN surrogate with the hidden representation learned by a pre-trained transformer (i.e., an open-source LLM), which significantly boosts its performance. These motivate us to propose our INSTruction optimization usIng Neural bandits Coupled with Transformers} (INSTINCT) algorithm. We perform instruction optimization for ChatGPT and use extensive experiments to show that our INSTINCT consistently outperforms the existing methods in different tasks, such as in various instruction induction tasks and the task of improving the zero-shot chain-of-thought instruction.
Boosting Dermatoscopic Lesion Segmentation via Diffusion Models with Visual and Textual Prompts
Authors: Shiyi Du, Xiaosong Wang, Yongyi Lu, Yuyin Zhou, Shaoting Zhang, Alan Yuille, Kang Li, Zongwei Zhou
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Abstract
Image synthesis approaches, e.g., generative adversarial networks, have been popular as a form of data augmentation in medical image analysis tasks. It is primarily beneficial to overcome the shortage of publicly accessible data and associated quality annotations. However, the current techniques often lack control over the detailed contents in generated images, e.g., the type of disease patterns, the location of lesions, and attributes of the diagnosis. In this work, we adapt the latest advance in the generative model, i.e., the diffusion model, with the added control flow using lesion-specific visual and textual prompts for generating dermatoscopic images. We further demonstrate the advantage of our diffusion model-based framework over the classical generation models in both the image quality and boosting the segmentation performance on skin lesions. It can achieve a 9% increase in the SSIM image quality measure and an over 5% increase in Dice coefficients over the prior arts.
Online Constraint Tightening in Stochastic Model Predictive Control: A Regression Approach
Authors: Alexandre Capone, Tim Brüdigam, Sandra Hirche
Subjects: Systems and Control (eess.SY); Machine Learning (cs.LG); Machine Learning (stat.ML)
Abstract
Solving chance-constrained stochastic optimal control problems is a significant challenge in control. This is because no analytical solutions exist for up to a handful of special cases. A common and computationally efficient approach for tackling chance-constrained stochastic optimal control problems consists of reformulating the chance constraints as hard constraints with a constraint-tightening parameter. However, in such approaches, the choice of constraint-tightening parameter remains challenging, and guarantees can mostly be obtained assuming that the process noise distribution is known a priori. Moreover, the chance constraints are often not tightly satisfied, leading to unnecessarily high costs. This work proposes a data-driven approach for learning the constraint-tightening parameters online during control. To this end, we reformulate the choice of constraint-tightening parameter for the closed-loop as a binary regression problem. We then leverage a highly expressive \gls{gp} model for binary regression to approximate the smallest constraint-tightening parameters that satisfy the chance constraints. By tuning the algorithm parameters appropriately, we show that the resulting constraint-tightening parameters satisfy the chance constraints up to an arbitrarily small margin with high probability. Our approach yields constraint-tightening parameters that tightly satisfy the chance constraints in numerical experiments, resulting in a lower average cost than three other state-of-the-art approaches.
Proximal Policy Optimization-Based Reinforcement Learning Approach for DC-DC Boost Converter Control: A Comparative Evaluation Against Traditional Control Techniques
Abstract
This article presents a proximal policy optimization (PPO) based reinforcement learning (RL) approach for DC-DC boost converter control, which is compared to traditional control methods. The performance of the PPO algorithm is evaluated using MATLAB Simulink co-simulation, and the results demonstrate that the most efficient approach for achieving short settling time and stability is to combine the PPO algorithm with reinforcement learning based control method. The simulation results indicate that the step response characteristics provided using the control method based on RL with the PPO algorithm outperform traditional control approaches, which can be used to improve DC-DC boost converter control. This research also highlights the inherent capability of the reinforcement learning method to enhance the performance of boost converter control.
Fast, Expressive SE$(n)$ Equivariant Networks through Weight-Sharing in Position-Orientation Space
Authors: Erik J Bekkers, Sharvaree Vadgama, Rob D Hesselink, Putri A van der Linden, David W Romero
Subjects: Machine Learning (cs.LG); Group Theory (math.GR)
Abstract
Based on the theory of homogeneous spaces we derive \textit{geometrically optimal edge attributes} to be used within the flexible message passing framework. We formalize the notion of weight sharing in convolutional networks as the sharing of message functions over point-pairs that should be treated equally. We define equivalence classes of point-pairs that are identical up to a transformation in the group and derive attributes that uniquely identify these classes. Weight sharing is then obtained by conditioning message functions on these attributes. As an application of the theory, we develop an efficient equivariant group convolutional network for processing 3D point clouds. The theory of homogeneous spaces tells us how to do group convolutions with feature maps over the homogeneous space of positions $\mathbb{R}^3$, position and orientations $\mathbb{R}^3 {\times} S^2$, and the group SE$(3)$ itself. Among these, $\mathbb{R}^3 {\times} S^2$ is an optimal choice due to the ability to represent directional information, which $\mathbb{R}^3$ methods cannot, and it significantly enhances computational efficiency compared to indexing features on the full SE$(3)$ group. We empirically support this claim by reaching state-of-the-art results -- in accuracy and speed -- on three different benchmarks: interatomic potential energy prediction, trajectory forecasting in N-body systems, and generating molecules via equivariant diffusion models.
T$^3$Bench: Benchmarking Current Progress in Text-to-3D Generation
Authors: Yuze He, Yushi Bai, Matthieu Lin, Wang Zhao, Yubin Hu, Jenny Sheng, Ran Yi, Juanzi Li, Yong-Jin Liu
Subjects: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL); Machine Learning (cs.LG)
Abstract
Recent methods in text-to-3D leverage powerful pretrained diffusion models to optimize NeRF. Notably, these methods are able to produce high-quality 3D scenes without training on 3D data. Due to the open-ended nature of the task, most studies evaluate their results with subjective case studies and user experiments, thereby presenting a challenge in quantitatively addressing the question: How has current progress in Text-to-3D gone so far? In this paper, we introduce T$^3$Bench, the first comprehensive text-to-3D benchmark containing diverse text prompts of three increasing complexity levels that are specially designed for 3D generation. To assess both the subjective quality and the text alignment, we propose two automatic metrics based on multi-view images produced by the 3D contents. The quality metric combines multi-view text-image scores and regional convolution to detect quality and view inconsistency. The alignment metric uses multi-view captioning and Large Language Model (LLM) evaluation to measure text-3D consistency. Both metrics closely correlate with different dimensions of human judgments, providing a paradigm for efficiently evaluating text-to-3D models. The benchmarking results, shown in Fig. 1, reveal performance differences among six prevalent text-to-3D methods. Our analysis further highlights the common struggles for current methods on generating surroundings and multi-object scenes, as well as the bottleneck of leveraging 2D guidance for 3D generation. Our project page is available at: https://t3bench.com.
Never Train from Scratch: Fair Comparison of Long-Sequence Models Requires Data-Driven Priors
Authors: Ido Amos, Jonathan Berant, Ankit Gupta
Subjects: Machine Learning (cs.LG); Computation and Language (cs.CL)
Abstract
Modeling long-range dependencies across sequences is a longstanding goal in machine learning and has led to architectures, such as state space models, that dramatically outperform Transformers on long sequences. However, these impressive empirical gains have been by and large demonstrated on benchmarks (e.g. Long Range Arena), where models are randomly initialized and trained to predict a target label from an input sequence. In this work, we show that random initialization leads to gross overestimation of the differences between architectures and that pretraining with standard denoising objectives, using $\textit{only the downstream task data}$, leads to dramatic gains across multiple architectures and to very small gaps between Transformers and state space models (SSMs). In stark contrast to prior works, we find vanilla Transformers to match the performance of S4 on Long Range Arena when properly pretrained, and we improve the best reported results of SSMs on the PathX-256 task by 20 absolute points. Subsequently, we analyze the utility of previously-proposed structured parameterizations for SSMs and show they become mostly redundant in the presence of data-driven initialization obtained through pretraining. Our work shows that, when evaluating different architectures on supervised tasks, incorporation of data-driven priors via pretraining is essential for reliable performance estimation, and can be done efficiently.
Exploring API Capabilities with Fieldwire
Authors: Nwosu Obinnaya Chikezie Victor
Subjects: Software Engineering (cs.SE); Systems and Control (eess.SY)
Abstract
Fieldwire, a cloud-based construction management software, has become a pivotal tool in the construction industry. It offers a comprehensive suite of features encompassing project management, task tracking, document management, and collaboration. With the rise of Application Programming Interfaces (APIs) in the software industry, Fieldwire has harnessed this trend to further empower construction professionals. APIs act as bridges between different software systems, and in Fieldwire's context, they hold the potential to integrate with specialized construction tools, eliminating data silos, manual data entry, and real-time information-sharing issues. This integration promises a streamlined and efficient construction management process, saving both time and resources. The research outlined in these abstract focuses on understanding Fieldwire's API capabilities, exploring integration possibilities with various construction tools, evaluating the impact of integration on efficiency and error reduction, establishing best practices, and offering recommendations to construction professionals. Python programming scripts are employed to visualize the benefits of API integration. Empirical findings indicate that Fieldwire's API significantly improves data accuracy, reduces project completion times by an average of 20%, and garners high user satisfaction. Such results are paramount in an industry reliant on precise data and efficient communication. This research underscores the transformative potential of Fieldwire's API and its relevance in modern construction management. It encourages construction professionals to embrace API integration for enhanced project outcomes and serves as an inspiration for software developers to innovate further in construction technology. As the construction industry evolves, API integration remains crucial for staying competitive and efficient.
ECoFLaP: Efficient Coarse-to-Fine Layer-Wise Pruning for Vision-Language Models
Authors: Yi-Lin Sung, Jaehong Yoon, Mohit Bansal
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
Abstract
Large Vision-Language Models (LVLMs) can understand the world comprehensively by integrating rich information from different modalities, achieving remarkable performance improvements on various multimodal downstream tasks. However, deploying LVLMs is often problematic due to their massive computational/energy costs and carbon consumption. Such issues make it infeasible to adopt conventional iterative global pruning, which is costly due to computing the Hessian matrix of the entire large model for sparsification. Alternatively, several studies have recently proposed layer-wise pruning approaches to avoid the expensive computation of global pruning and efficiently compress model weights according to their importance within a layer. However, these methods often suffer from suboptimal model compression due to their lack of a global perspective. To address this limitation in recent efficient pruning methods for large models, we propose Efficient Coarse-to-Fine Layer-Wise Pruning (ECoFLaP), a two-stage coarse-to-fine weight pruning approach for LVLMs. We first determine the sparsity ratios of different layers or blocks by leveraging the global importance score, which is efficiently computed based on the zeroth-order approximation of the global model gradients. Then, the multimodal model performs local layer-wise unstructured weight pruning based on globally-informed sparsity ratios. We validate our proposed method across various multimodal and unimodal models and datasets, demonstrating significant performance improvements over prevalent pruning techniques in the high-sparsity regime.
From Words to Watts: Benchmarking the Energy Costs of Large Language Model Inference
Authors: Siddharth Samsi, Dan Zhao, Joseph McDonald, Baolin Li, Adam Michaleas, Michael Jones, William Bergeron, Jeremy Kepner, Devesh Tiwari, Vijay Gadepally
Subjects: Computation and Language (cs.CL); Distributed, Parallel, and Cluster Computing (cs.DC)
Abstract
Large language models (LLMs) have exploded in popularity due to their new generative capabilities that go far beyond prior state-of-the-art. These technologies are increasingly being leveraged in various domains such as law, finance, and medicine. However, these models carry significant computational challenges, especially the compute and energy costs required for inference. Inference energy costs already receive less attention than the energy costs of training LLMs -- despite how often these large models are called on to conduct inference in reality (e.g., ChatGPT). As these state-of-the-art LLMs see increasing usage and deployment in various domains, a better understanding of their resource utilization is crucial for cost-savings, scaling performance, efficient hardware usage, and optimal inference strategies. In this paper, we describe experiments conducted to study the computational and energy utilization of inference with LLMs. We benchmark and conduct a preliminary analysis of the inference performance and inference energy costs of different sizes of LLaMA -- a recent state-of-the-art LLM -- developed by Meta AI on two generations of popular GPUs (NVIDIA V100 \& A100) and two datasets (Alpaca and GSM8K) to reflect the diverse set of tasks/benchmarks for LLMs in research and practice. We present the results of multi-node, multi-GPU inference using model sharding across up to 32 GPUs. To our knowledge, our work is the one of the first to study LLM inference performance from the perspective of computational and energy resources at this scale.
Efficient-3DiM: Learning a Generalizable Single-image Novel-view Synthesizer in One Day
Authors: Yifan Jiang, Hao Tang, Jen-Hao Rick Chang, Liangchen Song, Zhangyang Wang, Liangliang Cao
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
The task of novel view synthesis aims to generate unseen perspectives of an object or scene from a limited set of input images. Nevertheless, synthesizing novel views from a single image still remains a significant challenge in the realm of computer vision. Previous approaches tackle this problem by adopting mesh prediction, multi-plain image construction, or more advanced techniques such as neural radiance fields. Recently, a pre-trained diffusion model that is specifically designed for 2D image synthesis has demonstrated its capability in producing photorealistic novel views, if sufficiently optimized on a 3D finetuning task. Although the fidelity and generalizability are greatly improved, training such a powerful diffusion model requires a vast volume of training data and model parameters, resulting in a notoriously long time and high computational costs. To tackle this issue, we propose Efficient-3DiM, a simple but effective framework to learn a single-image novel-view synthesizer. Motivated by our in-depth analysis of the inference process of diffusion models, we propose several pragmatic strategies to reduce the training overhead to a manageable scale, including a crafted timestep sampling strategy, a superior 3D feature extractor, and an enhanced training scheme. When combined, our framework is able to reduce the total training time from 10 days to less than 1 day, significantly accelerating the training process under the same computational platform (one instance with 8 Nvidia A100 GPUs). Comprehensive experiments are conducted to demonstrate the efficiency and generalizability of our proposed method.
Understanding In-Context Learning in Transformers and LLMs by Learning to Learn Discrete Functions
Authors: Satwik Bhattamishra, Arkil Patel, Phil Blunsom, Varun Kanade
Subjects: Machine Learning (cs.LG); Computation and Language (cs.CL)
Abstract
In order to understand the in-context learning phenomenon, recent works have adopted a stylized experimental framework and demonstrated that Transformers can learn gradient-based learning algorithms for various classes of real-valued functions. However, the limitations of Transformers in implementing learning algorithms, and their ability to learn other forms of algorithms are not well understood. Additionally, the degree to which these capabilities are confined to attention-based models is unclear. Furthermore, it remains to be seen whether the insights derived from these stylized settings can be extrapolated to pretrained Large Language Models (LLMs). In this work, we take a step towards answering these questions by demonstrating the following: (a) On a test-bed with a variety of Boolean function classes, we find that Transformers can nearly match the optimal learning algorithm for 'simpler' tasks, while their performance deteriorates on more 'complex' tasks. Additionally, we find that certain attention-free models perform (almost) identically to Transformers on a range of tasks. (b) When provided a teaching sequence, i.e. a set of examples that uniquely identifies a function in a class, we show that Transformers learn more sample-efficiently. Interestingly, our results show that Transformers can learn to implement two distinct algorithms to solve a single task, and can adaptively select the more sample-efficient algorithm depending on the sequence of in-context examples. (c) Lastly, we show that extant LLMs, e.g. LLaMA-2, GPT-4, can compete with nearest-neighbor baselines on prediction tasks that are guaranteed to not be in their training set.
Human-oriented Representation Learning for Robotic Manipulation
Authors: Mingxiao Huo, Mingyu Ding, Chenfeng Xu, Thomas Tian, Xinghao Zhu, Yao Mu, Lingfeng Sun, Masayoshi Tomizuka, Wei Zhan
Abstract
Humans inherently possess generalizable visual representations that empower them to efficiently explore and interact with the environments in manipulation tasks. We advocate that such a representation automatically arises from simultaneously learning about multiple simple perceptual skills that are critical for everyday scenarios (e.g., hand detection, state estimate, etc.) and is better suited for learning robot manipulation policies compared to current state-of-the-art visual representations purely based on self-supervised objectives. We formalize this idea through the lens of human-oriented multi-task fine-tuning on top of pre-trained visual encoders, where each task is a perceptual skill tied to human-environment interactions. We introduce Task Fusion Decoder as a plug-and-play embedding translator that utilizes the underlying relationships among these perceptual skills to guide the representation learning towards encoding meaningful structure for what's important for all perceptual skills, ultimately empowering learning of downstream robotic manipulation tasks. Extensive experiments across a range of robotic tasks and embodiments, in both simulations and real-world environments, show that our Task Fusion Decoder consistently improves the representation of three state-of-the-art visual encoders including R3M, MVP, and EgoVLP, for downstream manipulation policy-learning. Project page: https://sites.google.com/view/human-oriented-robot-learning
Keyword: faster
SE(3)-Stochastic Flow Matching for Protein Backbone Generation
Authors: Avishek Joey Bose, Tara Akhound-Sadegh, Kilian Fatras, Guillaume Huguet, Jarrid Rector-Brooks, Cheng-Hao Liu, Andrei Cristian Nica, Maksym Korablyov, Michael Bronstein, Alexander Tong
Abstract
The computational design of novel protein structures has the potential to impact numerous scientific disciplines greatly. Toward this goal, we introduce $\text{FoldFlow}$ a series of novel generative models of increasing modeling power based on the flow-matching paradigm over $3\text{D}$ rigid motions -- i.e. the group $\text{SE(3)}$ -- enabling accurate modeling of protein backbones. We first introduce $\text{FoldFlow-Base}$, a simulation-free approach to learning deterministic continuous-time dynamics and matching invariant target distributions on $\text{SE(3)}$. We next accelerate training by incorporating Riemannian optimal transport to create $\text{FoldFlow-OT}$, leading to the construction of both more simple and stable flows. Finally, we design $\text{FoldFlow-SFM}$ coupling both Riemannian OT and simulation-free training to learn stochastic continuous-time dynamics over $\text{SE(3)}$. Our family of $\text{FoldFlow}$ generative models offer several key advantages over previous approaches to the generative modeling of proteins: they are more stable and faster to train than diffusion-based approaches, and our models enjoy the ability to map any invariant source distribution to any invariant target distribution over $\text{SE(3)}$. Empirically, we validate our FoldFlow models on protein backbone generation of up to $300$ amino acids leading to high-quality designable, diverse, and novel samples.
Practical, Private Assurance of the Value of Collaboration
Authors: Hassan Jameel Asghar, Zhigang Lu, Zhongrui Zhao, Dali Kaafar
Subjects: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
Abstract
Two parties wish to collaborate on their datasets. However, before they reveal their datasets to each other, the parties want to have the guarantee that the collaboration would be fruitful. We look at this problem from the point of view of machine learning, where one party is promised an improvement on its prediction model by incorporating data from the other party. The parties would only wish to collaborate further if the updated model shows an improvement in accuracy. Before this is ascertained, the two parties would not want to disclose their models and datasets. In this work, we construct an interactive protocol for this problem based on the fully homomorphic encryption scheme over the Torus (TFHE) and label differential privacy, where the underlying machine learning model is a neural network. Label differential privacy is used to ensure that computations are not done entirely in the encrypted domain, which is a significant bottleneck for neural network training according to the current state-of-the-art FHE implementations. We prove the security of our scheme in the universal composability framework assuming honest-but-curious parties, but where one party may not have any expertise in labelling its initial dataset. Experiments show that we can obtain the output, i.e., the accuracy of the updated model, with time many orders of magnitude faster than a protocol using entirely FHE operations.
A Faster Deterministic Approximation Algorithm for TTP-2
Authors: Yuga Kanaya, Kenjiro Takazawa
Subjects: Data Structures and Algorithms (cs.DS); Combinatorics (math.CO)
Abstract
The traveling tournament problem (TTP) is to minimize the total traveling distance of all teams in a double round-robin tournament. In this paper, we focus on TTP-2, in which each team plays at most two consecutive home games and at most two consecutive away games. For the case where the number of teams $n\equiv2$ (mod 4), Zhao and Xiao (2022) presented a $(1+5/n)$-approximation algorithm. This is a randomized algorithm running in $O(n^3)$ time, and its derandomized version runs in $O(n^4)$ time. In this paper, we present a faster deterministic algorithm running in $O(n^3)$ time, with approximation ratio $1+9/n$. This ratio improves the previous approximation ratios of the deterministic algorithms with the same time complexity.
On the Last-iterate Convergence in Time-varying Zero-sum Games: Extra Gradient Succeeds where Optimism Fails
Authors: Yi Feng, Hu Fu, Qun Hu, Ping Li, Ioannis Panageas, Bo Peng, Xiao Wang
Subjects: Computer Science and Game Theory (cs.GT)
Abstract
Last-iterate convergence has received extensive study in two player zero-sum games starting from bilinear, convex-concave up to settings that satisfy the MVI condition. Typical methods that exhibit last-iterate convergence for the aforementioned games include extra-gradient (EG) and optimistic gradient descent ascent (OGDA). However, all the established last-iterate convergence results hold for the restrictive setting where the underlying repeated game does not change over time. Recently, a line of research has focused on regret analysis of OGDA in time-varying games, i.e., games where payoffs evolve with time; the last-iterate behavior of OGDA and EG in time-varying environments remains unclear though. In this paper, we study the last-iterate behavior of various algorithms in two types of unconstrained, time-varying, bilinear zero-sum games: periodic and convergent perturbed games. These models expand upon the usual repeated game formulation and incorporate external environmental factors, such as the seasonal effects on species competition and vanishing external noise. In periodic games, we prove that EG will converge while OGDA and momentum method will diverge. This is quite surprising, as to the best of our knowledge, it is the first result that indicates EG and OGDA have qualitatively different last-iterate behaviors and do not exhibit similar behavior. In convergent perturbed games, we prove all these algorithms converge as long as the game itself stabilizes with a faster rate than $1/t$.
Foundation Reinforcement Learning: towards Embodied Generalist Agents with Foundation Prior Assistance
Authors: Weirui Ye, Yunsheng Zhang, Mengchen Wang, Shengjie Wang, Xianfan Gu, Pieter Abbeel, Yang Gao
Abstract
Recently, people have shown that large-scale pre-training from internet-scale data is the key to building generalist models, as witnessed in NLP. To build embodied generalist agents, we and many other researchers hypothesize that such foundation prior is also an indispensable component. However, it is unclear what is the proper concrete form to represent those embodied foundation priors and how they should be used in the downstream task. In this paper, we propose an intuitive and effective set of embodied priors that consist of foundation policy, value, and success reward. The proposed priors are based on the goal-conditioned MDP. To verify their effectiveness, we instantiate an actor-critic method assisted by the priors, called Foundation Actor-Critic (FAC). We name our framework as Foundation Reinforcement Learning (FRL), since it completely relies on embodied foundation priors to explore, learn and reinforce. The benefits of FRL are threefold. (1) Sample efficient. With foundation priors, FAC learns significantly faster than traditional RL. Our evaluation on the Meta-World has proved that FAC can achieve 100% success rates for 7/8 tasks under less than 200k frames, which outperforms the baseline method with careful manual-designed rewards under 1M frames. (2) Robust to noisy priors. Our method tolerates the unavoidable noise in embodied foundation models. We show that FAC works well even under heavy noise or quantization errors. (3) Minimal human intervention: FAC completely learns from the foundation priors, without the need of human-specified dense reward, or providing teleoperated demos. Thus, FAC can be easily scaled up. We believe our FRL framework could enable the future robot to autonomously explore and learn without human intervention in the physical world. In summary, our proposed FRL is a novel and powerful learning paradigm, towards achieving embodied generalist agents.
ED-NeRF: Efficient Text-Guided Editing of 3D Scene using Latent Space NeRF
Abstract
Recently, there has been a significant advancement in text-to-image diffusion models, leading to groundbreaking performance in 2D image generation. These advancements have been extended to 3D models, enabling the generation of novel 3D objects from textual descriptions. This has evolved into NeRF editing methods, which allow the manipulation of existing 3D objects through textual conditioning. However, existing NeRF editing techniques have faced limitations in their performance due to slow training speeds and the use of loss functions that do not adequately consider editing. To address this, here we present a novel 3D NeRF editing approach dubbed ED-NeRF by successfully embedding real-world scenes into the latent space of the latent diffusion model (LDM) through a unique refinement layer. This approach enables us to obtain a NeRF backbone that is not only faster but also more amenable to editing compared to traditional image space NeRF editing. Furthermore, we propose an improved loss function tailored for editing by migrating the delta denoising score (DDS) distillation loss, originally used in 2D image editing to the three-dimensional domain. This novel loss function surpasses the well-known score distillation sampling (SDS) loss in terms of suitability for editing purposes. Our experimental results demonstrate that ED-NeRF achieves faster editing speed while producing improved output quality compared to state-of-the-art 3D editing models.
Stability Improvements for Fast Matrix Multiplication
Abstract
We implement an Augmented Lagrangian method to minimize a constrained least-squares cost function designed to find polyadic decompositions of the matrix multiplication tensor. We use this method to obtain new discrete decompositions and parameter families of decompositions. Using these parametrizations, faster and more stable matrix multiplication algorithms can be discovered.
FroSSL: Frobenius Norm Minimization for Self-Supervised Learning
Authors: Oscar Skean, Aayush Dhakal, Nathan Jacobs, Luis Gonzalo Sanchez Giraldo
Abstract
Self-supervised learning (SSL) is an increasingly popular paradigm for representation learning. Recent methods can be classified as sample-contrastive, dimension-contrastive, or asymmetric network-based, with each family having its own approach to avoiding informational collapse. While dimension-contrastive methods converge to similar solutions as sample-contrastive methods, it can be empirically shown that some methods require more epochs of training to converge. Motivated by closing this divide, we present the objective function FroSSL which is both sample- and dimension-contrastive up to embedding normalization. FroSSL works by minimizing covariance Frobenius norms for avoiding collapse and minimizing mean-squared error for augmentation invariance. We show that FroSSL converges more quickly than a variety of other SSL methods and provide theoretical and empirical support that this faster convergence is due to how FroSSL affects the eigenvalues of the embedding covariance matrices. We also show that FroSSL learns competitive representations on linear probe evaluation when used to train a ResNet18 on the CIFAR-10, CIFAR-100, STL-10, and ImageNet datasets.
SemiReward: A General Reward Model for Semi-supervised Learning
Authors: Siyuan Li, Weiyang Jin, Zedong Wang, Fang Wu, Zicheng Liu, Cheng Tan, Stan Z. Li
Abstract
Semi-supervised learning (SSL) has witnessed great progress with various improvements in the self-training framework with pseudo labeling. The main challenge is how to distinguish high-quality pseudo labels against the confirmation bias. However, existing pseudo-label selection strategies are limited to pre-defined schemes or complex hand-crafted policies specially designed for classification, failing to achieve high-quality labels, fast convergence, and task versatility simultaneously. To these ends, we propose a Semi-supervised Reward framework (SemiReward) that predicts reward scores to evaluate and filter out high-quality pseudo labels, which is pluggable to mainstream SSL methods in wide task types and scenarios. To mitigate confirmation bias, SemiReward is trained online in two stages with a generator model and subsampling strategy. With classification and regression tasks on 13 standard SSL benchmarks of three modalities, extensive experiments verify that SemiReward achieves significant performance gains and faster convergence speeds upon Pseudo Label, FlexMatch, and Free/SoftMatch.
Retrieval meets Long Context Large Language Models
Abstract
Extending the context window of large language models (LLMs) is getting popular recently, while the solution of augmenting LLMs with retrieval has existed for years. The natural questions are: i) Retrieval-augmentation versus long context window, which one is better for downstream tasks? ii) Can both methods be combined to get the best of both worlds? In this work, we answer these questions by studying both solutions using two state-of-the-art pretrained LLMs, i.e., a proprietary 43B GPT and LLaMA2-70B. Perhaps surprisingly, we find that LLM with 4K context window using simple retrieval-augmentation at generation can achieve comparable performance to finetuned LLM with 16K context window via positional interpolation on long context tasks, while taking much less computation. More importantly, we demonstrate that retrieval can significantly improve the performance of LLMs regardless of their extended context window sizes. Our best model, retrieval-augmented LLaMA2-70B with 32K context window, outperforms GPT-3.5-turbo-16k and Davinci003 in terms of average score on seven long context tasks including question answering and query-based summarization. It also outperforms its non-retrieval LLaMA2-70B-32k baseline by a margin, while being much faster at generation. Our study provides general insights on the choice of retrieval-augmentation versus long context extension of LLM for practitioners.
Keyword: mobile
Online Proactive Multi-Task Assignment with Resource Availability Anticipation
Abstract
With the emergence of services and online applications as taxi dispatching, crowdsourcing, package or food delivery, industrials and researchers are paying attention to the online multi-task assignment optimization field to quickly and efficiently met demands. In this context, this paper is interested in the multi-task assignment problem where multiple requests (e.g. tasks) arrive over time and must be dynamically matched to (mobile) agents. This optimization problem is known to be NP-hard. In order to treat this problem with a proactive mindset, we propose to use a receding-horizon approach to determine which resources (e.g. taxis, mobile agents, drones, robots) would be available within this (possibly dynamic) receding-horizon to meet the current set of requests (i.e. tasks) as good as possible. Contrarily to several works in this domain, we have chosen to make no assumption concerning future locations of requests. To achieve fast optimized online solutions in terms of costs and amount of allocated tasks, we have designed a genetic algorithm based on a fitness function integrating the traveled distance and the age of the requests. We compared our proactive multi-task assignment with resource availability anticipation approach with a classical reactive approach. The results obtained in two benchmark problems, one synthetic and another based on real data, show that our resource availability anticipation method can achieve better results in terms of costs (e.g. traveled distance) and amount of allocated tasks than reactive approaches while decreasing resources idle time.
Event-Enhanced Multi-Modal Spiking Neural Network for Dynamic Obstacle Avoidance
Authors: Yang Wang, Bo Dong, Yuji Zhang, Yunduo Zhou, Haiyang Mei, Ziqi Wei, Xin Yang
Abstract
Autonomous obstacle avoidance is of vital importance for an intelligent agent such as a mobile robot to navigate in its environment. Existing state-of-the-art methods train a spiking neural network (SNN) with deep reinforcement learning (DRL) to achieve energy-efficient and fast inference speed in complex/unknown scenes. These methods typically assume that the environment is static while the obstacles in real-world scenes are often dynamic. The movement of obstacles increases the complexity of the environment and poses a great challenge to the existing methods. In this work, we approach robust dynamic obstacle avoidance twofold. First, we introduce the neuromorphic vision sensor (i.e., event camera) to provide motion cues complementary to the traditional Laser depth data for handling dynamic obstacles. Second, we develop an DRL-based event-enhanced multimodal spiking actor network (EEM-SAN) that extracts information from motion events data via unsupervised representation learning and fuses Laser and event camera data with learnable thresholding. Experiments demonstrate that our EEM-SAN outperforms state-of-the-art obstacle avoidance methods by a significant margin, especially for dynamic obstacle avoidance.
Automatic Multi-Path Web Story Creation from a Structural Article
Authors: Daniel Nkemelu, Peggy Chi, Daniel Castro Chin, Krishna Srinivasan, Irfan Essa
Abstract
Web articles such as Wikipedia serve as one of the major sources of knowledge dissemination and online learning. However, their in-depth information--often in a dense text format--may not be suitable for mobile browsing, even in a responsive UI. We propose an automatic approach that converts a structural article of any length into a set of interactive Web Stories that are ideal for mobile experiences. We focused on Wikipedia articles and developed Wiki2Story, a pipeline based on language and layout models, to demonstrate the concept. Wiki2Story dynamically slices an article and plans one to multiple Story paths according to the document hierarchy. For each slice, it generates a multi-page summary Story composed of text and image pairs in visually-appealing layouts. We derived design principles from an analysis of manually-created Story practices. We executed our pipeline on 500 Wikipedia documents and conducted user studies to review selected outputs. Results showed that Wiki2Story effectively captured and presented salient content from the original articles and sparked interest in viewers.
ShaSTA-Fuse: Camera-LiDAR Sensor Fusion to Model Shape and Spatio-Temporal Affinities for 3D Multi-Object Tracking
Abstract
3D multi-object tracking (MOT) is essential for an autonomous mobile agent to safely navigate a scene. In order to maximize the perception capabilities of the autonomous agent, we aim to develop a 3D MOT framework that fuses camera and LiDAR sensor information. Building on our prior LiDAR-only work, ShaSTA, which models shape and spatio-temporal affinities for 3D MOT, we propose a novel camera-LiDAR fusion approach for learning affinities. At its core, this work proposes a fusion technique that generates a rich sensory signal incorporating information about depth and distant objects to enhance affinity estimation for improved data association, track lifecycle management, false-positive elimination, false-negative propagation, and track confidence score refinement. Our main contributions include a novel fusion approach for combining camera and LiDAR sensory signals to learn affinities, and a first-of-its-kind multimodal sequential track confidence refinement technique that fuses 2D and 3D detections. Additionally, we perform an ablative analysis on each fusion step to demonstrate the added benefits of incorporating the camera sensor, particular for small, distant objects that tend to suffer from the depth-sensing limits and sparsity of LiDAR sensors. In sum, our technique achieves state-of-the-art performance on the nuScenes benchmark amongst multimodal 3D MOT algorithms using CenterPoint detections.
Hybrid Inception Architecture with Residual Connection: Fine-tuned Inception-ResNet Deep Learning Model for Lung Inflammation Diagnosis from Chest Radiographs
Abstract
Diagnosing lung inflammation, particularly pneumonia, is of paramount importance for effectively treating and managing the disease. Pneumonia is a common respiratory infection caused by bacteria, viruses, or fungi and can indiscriminately affect people of all ages. As highlighted by the World Health Organization (WHO), this prevalent disease tragically accounts for a substantial 15% of global mortality in children under five years of age. This article presents a comparative study of the Inception-ResNet deep learning model's performance in diagnosing pneumonia from chest radiographs. The study leverages Mendeleys chest X-ray images dataset, which contains 5856 2D images, including both Viral and Bacterial Pneumonia X-ray images. The Inception-ResNet model is compared with seven other state-of-the-art convolutional neural networks (CNNs), and the experimental results demonstrate the Inception-ResNet model's superiority in extracting essential features and saving computation runtime. Furthermore, we examine the impact of transfer learning with fine-tuning in improving the performance of deep convolutional models. This study provides valuable insights into using deep learning models for pneumonia diagnosis and highlights the potential of the Inception-ResNet model in this field. In classification accuracy, Inception-ResNet-V2 showed superior performance compared to other models, including ResNet152V2, MobileNet-V3 (Large and Small), EfficientNetV2 (Large and Small), InceptionV3, and NASNet-Mobile, with substantial margins. It outperformed them by 2.6%, 6.5%, 7.1%, 13%, 16.1%, 3.9%, and 1.6%, respectively, demonstrating its significant advantage in accurate classification.
High order numerical methods based on quadratic spline collocation method and averaged L1 scheme for the variable-order time fractional mobile/immobile diffusion equation
Authors: Xiao Ye, Jun Liu, Bingyin Zhang, Hongfei Fu, Yue Liu
Abstract
In this paper, we consider the variable-order time fractional mobile/immobile diffusion (TF-MID) equation in two-dimensional spatial domain, where the fractional order $\alpha(t)$ satisfies $0<\alpha_{}\leq \alpha(t)\leq \alpha^{}<1$. We combine the quadratic spline collocation (QSC) method and the $L1^+$ formula to propose a QSC-$L1^+$ scheme. It can be proved that, the QSC-$L1^+$ scheme is unconditionally stable and convergent with $\mathcal{O}(\tau^{\min{{3-\alpha^*-\alpha(0),2}}} + \Delta x^{2}+\Delta y^{2})$, where $\tau$, $\Delta x$ and $\Delta y$ are the temporal and spatial step sizes, respectively. With some proper assumptions on $\alpha(t)$, the QSC-$L1^+$ scheme has second temporal convergence order even on the uniform mesh, without any restrictions on the solution of the equation. We further construct a novel alternating direction implicit (ADI) framework to develop an ADI-QSC-$L1^+$ scheme, which has the same unconditionally stability and convergence orders. In addition, a fast implementation for the ADI-QSC-$L1^+$ scheme based on the exponential-sum-approximation (ESA) technique is proposed. Moreover, we also introduce the optimal QSC method to improve the spatial convergence to fourth-order. Numerical experiments are attached to support the theoretical analysis, and to demonstrate the effectiveness of the proposed schemes.
R-LGP: A Reachability-guided Logic-geometric Programming Framework for Optimal Task and Motion Planning on Mobile Manipulators
Authors: Kim Tien Ly, Valeriy Semenov, Mattia Risiglione, Wolfgang Merkt, Ioannis Havoutis
Abstract
This paper presents an optimization-based solution to task and motion planning (TAMP) on mobile manipulators. Logic-geometric programming (LGP) has shown promising capabilities for optimally dealing with hybrid TAMP problems that involve abstract and geometric constraints. However, LGP does not scale well to high-dimensional systems (e.g. mobile manipulators) and can suffer from obstacle avoidance issues. In this work, we extend LGP with a sampling-based reachability graph to enable solving optimal TAMP on high-DoF mobile manipulators. The proposed reachability graph can incorporate environmental information (obstacles) to provide the planner with sufficient geometric constraints. This reachability-aware heuristic efficiently prunes infeasible sequences of actions in the continuous domain, hence, it reduces replanning by securing feasibility at the final full trajectory optimization. Our framework proves to be time-efficient in computing optimal and collision-free solutions, while outperforming the current state of the art on metrics of success rate, planning time, path length and number of steps. We validate our framework on the physical Toyota HSR robot and report comparisons on a series of mobile manipulation tasks of increasing difficulty.
Keyword: pruning
Junk DNA Hypothesis: A Task-Centric Angle of LLM Pre-trained Weights through Sparsity
Authors: Lu Yin, Shiwei Liu, Ajay Jaiswal, Souvik Kundu, Zhangyang Wang
Abstract
The traditional notion of "Junk DNA" has long been linked to non-coding segments within the human genome, constituting roughly 98% of its composition. However, recent research has unveiled the critical roles some of these seemingly non-functional DNA sequences play in cellular processes. Intriguingly, the weights within deep neural networks exhibit a remarkable similarity to the redundancy observed in human genes. It was believed that weights in gigantic models contained excessive redundancy, and could be removed without compromising performance. This paper challenges this conventional wisdom by presenting a compelling counter-argument. We employ sparsity as a tool to isolate and quantify the nuanced significance of low-magnitude weights in pre-trained large language models (LLMs). Our study demonstrates a strong correlation between these weight magnitudes and the knowledge they encapsulate, from a downstream task-centric angle. we raise the "Junk DNA Hypothesis" backed by our in-depth investigation: while small-magnitude weights may appear "useless" for simple tasks and suitable for pruning, they actually encode crucial knowledge necessary for solving more difficult downstream tasks. Removing these seemingly insignificant weights can lead to irreversible knowledge forgetting and performance damage in difficult tasks. These findings offer fresh insights into how LLMs encode knowledge in a task-sensitive manner, pave future research direction in model pruning, and open avenues for task-aware conditional computation during inference.
Feather: An Elegant Solution to Effective DNN Sparsification
Authors: Athanasios Glentis Georgoulakis, George Retsinas, Petros Maragos
Abstract
Neural Network pruning is an increasingly popular way for producing compact and efficient models, suitable for resource-limited environments, while preserving high performance. While the pruning can be performed using a multi-cycle training and fine-tuning process, the recent trend is to encompass the sparsification process during the standard course of training. To this end, we introduce Feather, an efficient sparse training module utilizing the powerful Straight-Through Estimator as its core, coupled with a new thresholding operator and a gradient scaling technique, enabling robust, out-of-the-box sparsification performance. Feather's effectiveness and adaptability is demonstrated using various architectures on the CIFAR dataset, while on ImageNet it achieves state-of-the-art Top-1 validation accuracy using the ResNet-50 architecture, surpassing existing methods, including more complex and computationally heavy ones, by a considerable margin. Code is publicly available at https://github.com/athglentis/feather .
Convergence Analysis and Latency Minimization for Semi-Federated Learning in Massive IoT Networks
Authors: Jianyang Ren, Wanli Ni, Hui Tian, Gaofeng Nie
Subjects: Information Theory (cs.IT); Signal Processing (eess.SP)
Abstract
As the number of sensors becomes massive in Internet of Things (IoT) networks, the amount of data is humongous. To process data in real-time while protecting user privacy, federated learning (FL) has been regarded as an enabling technique to push edge intelligence into IoT networks with massive devices. However, FL latency increases dramatically due to the increase of the number of parameters in deep neural network and the limited computation and communication capabilities of IoT devices. To address this issue, we propose a semi-federated learning (SemiFL) paradigm in which network pruning and over-the-air computation are efficiently applied. To be specific, each small base station collects the raw data from its served sensors and trains its local pruned model. After that, the global aggregation of local gradients is achieved through over-the-air computation. We first analyze the performance of the proposed SemiFL by deriving its convergence upper bound. To reduce latency, a convergence-constrained SemiFL latency minimization problem is formulated. By decoupling the original problem into several sub-problems, iterative algorithms are designed to solve them efficiently. Finally, numerical simulations are conducted to verify the effectiveness of our proposed scheme in reducing latency and guaranteeing the identification accuracy.
Top-k contrast order-preserving pattern mining for time series classification
Authors: Youxi Wu, Yufei Meng, Yan Li, Lei Guo, Xingquan Zhu, Philippe Fournier-Viger, Xindong Wu
Abstract
Recently, order-preserving pattern (OPP) mining, a new sequential pattern mining method, has been proposed to mine frequent relative orders in a time series. Although frequent relative orders can be used as features to classify a time series, the mined patterns do not reflect the differences between two classes of time series well. To effectively discover the differences between time series, this paper addresses the top-k contrast OPP (COPP) mining and proposes a COPP-Miner algorithm to discover the top-k contrast patterns as features for time series classification, avoiding the problem of improper parameter setting. COPP-Miner is composed of three parts: extreme point extraction to reduce the length of the original time series, forward mining, and reverse mining to discover COPPs. Forward mining contains three steps: group pattern fusion strategy to generate candidate patterns, the support rate calculation method to efficiently calculate the support of a pattern, and two pruning strategies to further prune candidate patterns. Reverse mining uses one pruning strategy to prune candidate patterns and consists of applying the same process as forward mining. Experimental results validate the efficiency of the proposed algorithm and show that top-k COPPs can be used as features to obtain a better classification performance.
Co-Optimizing Cache Partitioning and Multi-Core Task Scheduling: Exploit Cache Sensitivity or Not?
Authors: Binqi Sun, Debayan Roy, Tomasz Kloda, Andrea Bastoni, Rodolfo Pellizzoni, Marco Caccamo
Subjects: Hardware Architecture (cs.AR); Distributed, Parallel, and Cluster Computing (cs.DC); Operating Systems (cs.OS)
Abstract
Cache partitioning techniques have been successfully adopted to mitigate interference among concurrently executing real-time tasks on multi-core processors. Considering that the execution time of a cache-sensitive task strongly depends on the cache available for it to use, co-optimizing cache partitioning and task allocation improves the system's schedulability. In this paper, we propose a hybrid multi-layer design space exploration technique to solve this multi-resource management problem. We explore the interplay between cache partitioning and schedulability by systematically interleaving three optimization layers, viz., (i) in the outer layer, we perform a breadth-first search combined with proactive pruning for cache partitioning; (ii) in the middle layer, we exploit a first-fit heuristic for allocating tasks to cores; and (iii) in the inner layer, we use the well-known recurrence relation for the schedulability analysis of non-preemptive fixed-priority (NP-FP) tasks in a uniprocessor setting. Although our focus is on NP-FP scheduling, we evaluate the flexibility of our framework in supporting different scheduling policies (NP-EDF, P-EDF) by plugging in appropriate analysis methods in the inner layer. Experiments show that, compared to the state-of-the-art techniques, the proposed framework can improve the real-time schedulability of NP-FP task sets by an average of 15.2% with a maximum improvement of 233.6% (when tasks are highly cache-sensitive) and a minimum of 1.6% (when cache sensitivity is low). For such task sets, we found that clustering similar-period (or mutually compatible) tasks often leads to higher schedulability (on average 7.6%) than clustering by cache sensitivity. In our evaluation, the framework also achieves good results for preemptive and dynamic-priority scheduling policies.
ECoFLaP: Efficient Coarse-to-Fine Layer-Wise Pruning for Vision-Language Models
Authors: Yi-Lin Sung, Jaehong Yoon, Mohit Bansal
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
Abstract
Large Vision-Language Models (LVLMs) can understand the world comprehensively by integrating rich information from different modalities, achieving remarkable performance improvements on various multimodal downstream tasks. However, deploying LVLMs is often problematic due to their massive computational/energy costs and carbon consumption. Such issues make it infeasible to adopt conventional iterative global pruning, which is costly due to computing the Hessian matrix of the entire large model for sparsification. Alternatively, several studies have recently proposed layer-wise pruning approaches to avoid the expensive computation of global pruning and efficiently compress model weights according to their importance within a layer. However, these methods often suffer from suboptimal model compression due to their lack of a global perspective. To address this limitation in recent efficient pruning methods for large models, we propose Efficient Coarse-to-Fine Layer-Wise Pruning (ECoFLaP), a two-stage coarse-to-fine weight pruning approach for LVLMs. We first determine the sparsity ratios of different layers or blocks by leveraging the global importance score, which is efficiently computed based on the zeroth-order approximation of the global model gradients. Then, the multimodal model performs local layer-wise unstructured weight pruning based on globally-informed sparsity ratios. We validate our proposed method across various multimodal and unimodal models and datasets, demonstrating significant performance improvements over prevalent pruning techniques in the high-sparsity regime.
Keyword: diffusion
Consistency Trajectory Models: Learning Probability Flow ODE Trajectory of Diffusion
Abstract
Consistency Models (CM) (Song et al., 2023) accelerate score-based diffusion model sampling at the cost of sample quality but lack a natural way to trade-off quality for speed. To address this limitation, we propose Consistency Trajectory Model (CTM), a generalization encompassing CM and score-based models as special cases. CTM trains a single neural network that can -- in a single forward pass -- output scores (i.e., gradients of log-density) and enables unrestricted traversal between any initial and final time along the Probability Flow Ordinary Differential Equation (ODE) in a diffusion process. CTM enables the efficient combination of adversarial training and denoising score matching loss to enhance performance and achieves new state-of-the-art FIDs for single-step diffusion model sampling on CIFAR-10 (FID 1.73) and ImageNet at 64X64 resolution (FID 2.06). CTM also enables a new family of sampling schemes, both deterministic and stochastic, involving long jumps along the ODE solution trajectories. It consistently improves sample quality as computational budgets increase, avoiding the degradation seen in CM. Furthermore, CTM's access to the score accommodates all diffusion model inference techniques, including exact likelihood computation.
Stochastic force inference via density estimation
Authors: Victor Chardès, Suryanarayana Maddu, Michael J. Shelley
Abstract
Inferring dynamical models from low-resolution temporal data continues to be a significant challenge in biophysics, especially within transcriptomics, where separating molecular programs from noise remains an important open problem. We explore a common scenario in which we have access to an adequate amount of cross-sectional samples at a few time-points, and assume that our samples are generated from a latent diffusion process. We propose an approach that relies on the probability flow associated with an underlying diffusion process to infer an autonomous, nonlinear force field interpolating between the distributions. Given a prior on the noise model, we employ score-matching to differentiate the force field from the intrinsic noise. Using relevant biophysical examples, we demonstrate that our approach can extract non-conservative forces from non-stationary data, that it learns equilibrium dynamics when applied to steady-state data, and that it can do so with both additive and multiplicative noise models.
SE(3)-Stochastic Flow Matching for Protein Backbone Generation
Authors: Avishek Joey Bose, Tara Akhound-Sadegh, Kilian Fatras, Guillaume Huguet, Jarrid Rector-Brooks, Cheng-Hao Liu, Andrei Cristian Nica, Maksym Korablyov, Michael Bronstein, Alexander Tong
Abstract
The computational design of novel protein structures has the potential to impact numerous scientific disciplines greatly. Toward this goal, we introduce $\text{FoldFlow}$ a series of novel generative models of increasing modeling power based on the flow-matching paradigm over $3\text{D}$ rigid motions -- i.e. the group $\text{SE(3)}$ -- enabling accurate modeling of protein backbones. We first introduce $\text{FoldFlow-Base}$, a simulation-free approach to learning deterministic continuous-time dynamics and matching invariant target distributions on $\text{SE(3)}$. We next accelerate training by incorporating Riemannian optimal transport to create $\text{FoldFlow-OT}$, leading to the construction of both more simple and stable flows. Finally, we design $\text{FoldFlow-SFM}$ coupling both Riemannian OT and simulation-free training to learn stochastic continuous-time dynamics over $\text{SE(3)}$. Our family of $\text{FoldFlow}$ generative models offer several key advantages over previous approaches to the generative modeling of proteins: they are more stable and faster to train than diffusion-based approaches, and our models enjoy the ability to map any invariant source distribution to any invariant target distribution over $\text{SE(3)}$. Empirically, we validate our FoldFlow models on protein backbone generation of up to $300$ amino acids leading to high-quality designable, diverse, and novel samples.
FT-Shield: A Watermark Against Unauthorized Fine-tuning in Text-to-Image Diffusion Models
Authors: Yingqian Cui, Jie Ren, Yuping Lin, Han Xu, Pengfei He, Yue Xing, Wenqi Fan, Hui Liu, Jiliang Tang
Subjects: Computer Vision and Pattern Recognition (cs.CV); Cryptography and Security (cs.CR)
Abstract
Text-to-image generative models based on latent diffusion models (LDM) have demonstrated their outstanding ability in generating high-quality and high-resolution images according to language prompt. Based on these powerful latent diffusion models, various fine-tuning methods have been proposed to achieve the personalization of text-to-image diffusion models such as artistic style adaptation and human face transfer. However, the unauthorized usage of data for model personalization has emerged as a prevalent concern in relation to copyright violations. For example, a malicious user may use the fine-tuning technique to generate images which mimic the style of a painter without his/her permission. In light of this concern, we have proposed FT-Shield, a watermarking approach specifically designed for the fine-tuning of text-to-image diffusion models to aid in detecting instances of infringement. We develop a novel algorithm for the generation of the watermark to ensure that the watermark on the training images can be quickly and accurately transferred to the generated images of text-to-image diffusion models. A watermark will be detected on an image by a binary watermark detector if the image is generated by a model that has been fine-tuned using the protected watermarked images. Comprehensive experiments were conducted to validate the effectiveness of FT-Shield.
EditVal: Benchmarking Diffusion Based Text-Guided Image Editing Methods
Abstract
A plethora of text-guided image editing methods have recently been developed by leveraging the impressive capabilities of large-scale diffusion-based generative models such as Imagen and Stable Diffusion. A standardized evaluation protocol, however, does not exist to compare methods across different types of fine-grained edits. To address this gap, we introduce EditVal, a standardized benchmark for quantitatively evaluating text-guided image editing methods. EditVal consists of a curated dataset of images, a set of editable attributes for each image drawn from 13 possible edit types, and an automated evaluation pipeline that uses pre-trained vision-language models to assess the fidelity of generated images for each edit type. We use EditVal to benchmark 8 cutting-edge diffusion-based editing methods including SINE, Imagic and Instruct-Pix2Pix. We complement this with a large-scale human study where we show that EditVall's automated evaluation pipeline is strongly correlated with human-preferences for the edit types we considered. From both the human study and automated evaluation, we find that: (i) Instruct-Pix2Pix, Null-Text and SINE are the top-performing methods averaged across different edit types, however {\it only} Instruct-Pix2Pix and Null-Text are able to preserve original image properties; (ii) Most of the editing methods fail at edits involving spatial operations (e.g., changing the position of an object). (iii) There is no `winner' method which ranks the best individually across a range of different edit types. We hope that our benchmark can pave the way to developing more reliable text-guided image editing tools in the future. We will publicly release EditVal, and all associated code and human-study templates to support these research directions in https://deep-ml-research.github.io/editval/.
Abstract
Diffusion models are a powerful class of generative models capable of mapping random noise in high-dimensional spaces to a target manifold through iterative denoising. In this work, we present a novel perspective on goal-conditioned reinforcement learning by framing it within the context of diffusion modeling. Analogous to the diffusion process, where Gaussian noise is used to create random trajectories that walk away from the data manifold, we construct trajectories that move away from potential goal states. We then learn a goal-conditioned policy analogous to the score function. This approach, which we call Merlin, can reach predefined or novel goals from an arbitrary initial state without learning a separate value function. We consider three choices for the noise model to replace Gaussian noise in diffusion - reverse play from the buffer, reverse dynamics model, and a novel non-parametric approach. We theoretically justify our approach and validate it on offline goal-reaching tasks. Empirical results are competitive with state-of-the-art methods, which suggests this perspective on diffusion for RL is a simple, scalable, and effective direction for sequential decision-making.
Ophiuchus: Scalable Modeling of Protein Structures through Hierarchical Coarse-graining SO(3)-Equivariant Autoencoders
Authors: Allan dos Santos Costa, Ilan Mitnikov, Mario Geiger, Manvitha Ponnapati, Tess Smidt, Joseph Jacobson
Abstract
Three-dimensional native states of natural proteins display recurring and hierarchical patterns. Yet, traditional graph-based modeling of protein structures is often limited to operate within a single fine-grained resolution, and lacks hourglass neural architectures to learn those high-level building blocks. We narrow this gap by introducing Ophiuchus, an SO(3)-equivariant coarse-graining model that efficiently operates on all heavy atoms of standard protein residues, while respecting their relevant symmetries. Our model departs from current approaches that employ graph modeling, instead focusing on local convolutional coarsening to model sequence-motif interactions in log-linear length complexity. We train Ophiuchus on contiguous fragments of PDB monomers, investigating its reconstruction capabilities across different compression rates. We examine the learned latent space and demonstrate its prompt usage in conformational interpolation, comparing interpolated trajectories to structure snapshots from the PDBFlex dataset. Finally, we leverage denoising diffusion probabilistic models (DDPM) to efficiently sample readily-decodable latent embeddings of diverse miniproteins. Our experiments demonstrate Ophiuchus to be a scalable basis for efficient protein modeling and generation.
MedDiffusion: Boosting Health Risk Prediction via Diffusion-based Data Augmentation
Abstract
Health risk prediction is one of the fundamental tasks under predictive modeling in the medical domain, which aims to forecast the potential health risks that patients may face in the future using their historical Electronic Health Records (EHR). Researchers have developed several risk prediction models to handle the unique challenges of EHR data, such as its sequential nature, high dimensionality, and inherent noise. These models have yielded impressive results. Nonetheless, a key issue undermining their effectiveness is data insufficiency. A variety of data generation and augmentation methods have been introduced to mitigate this issue by expanding the size of the training data set through the learning of underlying data distributions. However, the performance of these methods is often limited due to their task-unrelated design. To address these shortcomings, this paper introduces a novel, end-to-end diffusion-based risk prediction model, named MedDiffusion. It enhances risk prediction performance by creating synthetic patient data during training to enlarge sample space. Furthermore, MedDiffusion discerns hidden relationships between patient visits using a step-wise attention mechanism, enabling the model to automatically retain the most vital information for generating high-quality data. Experimental evaluation on four real-world medical datasets demonstrates that MedDiffusion outperforms 14 cutting-edge baselines in terms of PR-AUC, F1, and Cohen's Kappa. We also conduct ablation studies and benchmark our model against GAN-based alternatives to further validate the rationality and adaptability of our model design. Additionally, we analyze generated data to offer fresh insights into the model's interpretability.
Generalization in diffusion models arises from geometry-adaptive harmonic representation
Authors: Zahra Kadkhodaie, Florentin Guth, Eero P. Simoncelli, Stéphane Mallat
Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Abstract
High-quality samples generated with score-based reverse diffusion algorithms provide evidence that deep neural networks (DNN) trained for denoising can learn high-dimensional densities, despite the curse of dimensionality. However, recent reports of memorization of the training set raise the question of whether these networks are learning the "true" continuous density of the data. Here, we show that two denoising DNNs trained on non-overlapping subsets of a dataset learn nearly the same score function, and thus the same density, with a surprisingly small number of training images. This strong generalization demonstrates an alignment of powerful inductive biases in the DNN architecture and/or training algorithm with properties of the data distribution. We analyze these, demonstrating that the denoiser performs a shrinkage operation in a basis adapted to the underlying image. Examination of these bases reveals oscillating harmonic structures along contours and in homogeneous image regions. We show that trained denoisers are inductively biased towards these geometry-adaptive harmonic representations by demonstrating that they arise even when the network is trained on image classes such as low-dimensional manifolds, for which the harmonic basis is suboptimal. Additionally, we show that the denoising performance of the networks is near-optimal when trained on regular image classes for which the optimal basis is known to be geometry-adaptive and harmonic.
SweetDreamer: Aligning Geometric Priors in 2D Diffusion for Consistent Text-to-3D
Authors: Weiyu Li, Rui Chen, Xuelin Chen, Ping Tan
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
It is inherently ambiguous to lift 2D results from pre-trained diffusion models to a 3D world for text-to-3D generation. 2D diffusion models solely learn view-agnostic priors and thus lack 3D knowledge during the lifting, leading to the multi-view inconsistency problem. We find that this problem primarily stems from geometric inconsistency, and avoiding misplaced geometric structures substantially mitigates the problem in the final outputs. Therefore, we improve the consistency by aligning the 2D geometric priors in diffusion models with well-defined 3D shapes during the lifting, addressing the vast majority of the problem. This is achieved by fine-tuning the 2D diffusion model to be viewpoint-aware and to produce view-specific coordinate maps of canonically oriented 3D objects. In our process, only coarse 3D information is used for aligning. This "coarse" alignment not only resolves the multi-view inconsistency in geometries but also retains the ability in 2D diffusion models to generate detailed and diversified high-quality objects unseen in the 3D datasets. Furthermore, our aligned geometric priors (AGP) are generic and can be seamlessly integrated into various state-of-the-art pipelines, obtaining high generalizability in terms of unseen shapes and visual appearance while greatly alleviating the multi-view inconsistency problem. Our method represents a new state-of-the-art performance with an 85+% consistency rate by human evaluation, while many previous methods are around 30%. Our project page is https://sweetdreamer3d.github.io/
MagicDrive: Street View Generation with Diverse 3D Geometry Control
Abstract
Recent advancements in diffusion models have significantly enhanced the data synthesis with 2D control. Yet, precise 3D control in street view generation, crucial for 3D perception tasks, remains elusive. Specifically, utilizing Bird's-Eye View (BEV) as the primary condition often leads to challenges in geometry control (e.g., height), affecting the representation of object shapes, occlusion patterns, and road surface elevations, all of which are essential to perception data synthesis, especially for 3D object detection tasks. In this paper, we introduce MagicDrive, a novel street view generation framework offering diverse 3D geometry controls, including camera poses, road maps, and 3D bounding boxes, together with textual descriptions, achieved through tailored encoding strategies. Besides, our design incorporates a cross-view attention module, ensuring consistency across multiple camera views. With MagicDrive, we achieve high-fidelity street-view synthesis that captures nuanced 3D geometry and various scene descriptions, enhancing tasks like BEV segmentation and 3D object detection.
On Memorization in Diffusion Models
Authors: Xiangming Gu, Chao Du, Tianyu Pang, Chongxuan Li, Min Lin, Ye Wang
Abstract
Due to their capacity to generate novel and high-quality samples, diffusion models have attracted significant research interest in recent years. Notably, the typical training objective of diffusion models, i.e., denoising score matching, has a closed-form optimal solution that can only generate training data replicating samples. This indicates that a memorization behavior is theoretically expected, which contradicts the common generalization ability of state-of-the-art diffusion models, and thus calls for a deeper understanding. Looking into this, we first observe that memorization behaviors tend to occur on smaller-sized datasets, which motivates our definition of effective model memorization (EMM), a metric measuring the maximum size of training data at which a learned diffusion model approximates its theoretical optimum. Then, we quantify the impact of the influential factors on these memorization behaviors in terms of EMM, focusing primarily on data distribution, model configuration, and training procedure. Besides comprehensive empirical results identifying the influential factors, we surprisingly find that conditioning training data on uninformative random labels can significantly trigger the memorization in diffusion models. Our study holds practical significance for diffusion model users and offers clues to theoretical research in deep generative models. Code is available at https://github.com/sail-sg/DiffMemorize.
Abstract
We tackle the problem of sampling from intractable high-dimensional density functions, a fundamental task that often appears in machine learning and statistics. We extend recent sampling-based approaches that leverage controlled stochastic processes to model approximate samples from these target densities. The main drawback of these approaches is that the training objective requires full trajectories to compute, resulting in sluggish credit assignment issues due to use of entire trajectories and a learning signal present only at the terminal time. In this work, we present Diffusion Generative Flow Samplers (DGFS), a sampling-based framework where the learning process can be tractably broken down into short partial trajectory segments, via parameterizing an additional "flow function". Our method takes inspiration from the theory developed for generative flow networks (GFlowNets), allowing us to make use of intermediate learning signals and benefit from off-policy exploration capabilities. Through a variety of challenging experiments, we demonstrate that DGFS results in more accurate estimates of the normalization constant than closely-related prior methods.
Robust Tracking for a 3D Diffusion Equation: Controlling Seismicity Rate in Geothermal Reservoirs
Authors: Diego Gutierrez-Oribio, Ioannis Stefanou
Subjects: Systems and Control (eess.SY); Optimization and Control (math.OC)
Abstract
Deep Geothermal Energy has significant potential to meet the large-scale needs of the energy sector. However, the injection of fluids into the earth's crust, upon which it relies, can lead to the formation of new seismogenic faults or the reactivation of existing ones, thereby causing earthquakes. To date, no effective method exists for mitigating these human-induced earthquakes. In this study, we propose a novel approach based on control theory to address this issue. First, we model induced seismicity resulting from fluid injections in a geothermal reservoir using a diffusion equation in three dimensions. Then, we design a robust tracking control approach to force the seismicity rate to follow the desired references. In this way, the induced seismicity is minimized while ensuring fluid circulation for the needs of energy production. The designed control guarantees the stabilization of the error variable even in the presence of system uncertainties and unknown dynamics. Finally, we present simulations of a geothermal reservoir under different scenarios of intermittent energy demand to show the reliability and performance of the control approach, opening new perspectives for field experiments based on real-time regulators for the first time.
ED-NeRF: Efficient Text-Guided Editing of 3D Scene using Latent Space NeRF
Abstract
Recently, there has been a significant advancement in text-to-image diffusion models, leading to groundbreaking performance in 2D image generation. These advancements have been extended to 3D models, enabling the generation of novel 3D objects from textual descriptions. This has evolved into NeRF editing methods, which allow the manipulation of existing 3D objects through textual conditioning. However, existing NeRF editing techniques have faced limitations in their performance due to slow training speeds and the use of loss functions that do not adequately consider editing. To address this, here we present a novel 3D NeRF editing approach dubbed ED-NeRF by successfully embedding real-world scenes into the latent space of the latent diffusion model (LDM) through a unique refinement layer. This approach enables us to obtain a NeRF backbone that is not only faster but also more amenable to editing compared to traditional image space NeRF editing. Furthermore, we propose an improved loss function tailored for editing by migrating the delta denoising score (DDS) distillation loss, originally used in 2D image editing to the three-dimensional domain. This novel loss function surpasses the well-known score distillation sampling (SDS) loss in terms of suitability for editing purposes. Our experimental results demonstrate that ED-NeRF achieves faster editing speed while producing improved output quality compared to state-of-the-art 3D editing models.
High order numerical methods based on quadratic spline collocation method and averaged L1 scheme for the variable-order time fractional mobile/immobile diffusion equation
Authors: Xiao Ye, Jun Liu, Bingyin Zhang, Hongfei Fu, Yue Liu
Abstract
In this paper, we consider the variable-order time fractional mobile/immobile diffusion (TF-MID) equation in two-dimensional spatial domain, where the fractional order $\alpha(t)$ satisfies $0<\alpha_{}\leq \alpha(t)\leq \alpha^{}<1$. We combine the quadratic spline collocation (QSC) method and the $L1^+$ formula to propose a QSC-$L1^+$ scheme. It can be proved that, the QSC-$L1^+$ scheme is unconditionally stable and convergent with $\mathcal{O}(\tau^{\min{{3-\alpha^*-\alpha(0),2}}} + \Delta x^{2}+\Delta y^{2})$, where $\tau$, $\Delta x$ and $\Delta y$ are the temporal and spatial step sizes, respectively. With some proper assumptions on $\alpha(t)$, the QSC-$L1^+$ scheme has second temporal convergence order even on the uniform mesh, without any restrictions on the solution of the equation. We further construct a novel alternating direction implicit (ADI) framework to develop an ADI-QSC-$L1^+$ scheme, which has the same unconditionally stability and convergence orders. In addition, a fast implementation for the ADI-QSC-$L1^+$ scheme based on the exponential-sum-approximation (ESA) technique is proposed. Moreover, we also introduce the optimal QSC method to improve the spatial convergence to fourth-order. Numerical experiments are attached to support the theoretical analysis, and to demonstrate the effectiveness of the proposed schemes.
Magicremover: Tuning-free Text-guided Image inpainting with Diffusion Models
Authors: Siyuan Yang, Lu Zhang, Liqian Ma, Yu Liu, JingJing Fu, You He
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Image inpainting aims to fill in the missing pixels with visually coherent and semantically plausible content. Despite the great progress brought from deep generative models, this task still suffers from i. the difficulties in large-scale realistic data collection and costly model training; and ii. the intrinsic limitations in the traditionally user-defined binary masks on objects with unclear boundaries or transparent texture. In this paper, we propose MagicRemover, a tuning-free method that leverages the powerful diffusion models for text-guided image inpainting. We introduce an attention guidance strategy to constrain the sampling process of diffusion models, enabling the erasing of instructed areas and the restoration of occluded content. We further propose a classifier optimization algorithm to facilitate the denoising stability within less sampling steps. Extensive comparisons are conducted among our MagicRemover and state-of-the-art methods including quantitative evaluation and user study, demonstrating the significant improvement of MagicRemover on high-quality image inpainting. We will release our code at https://github.com/exisas/Magicremover.
Boosting Dermatoscopic Lesion Segmentation via Diffusion Models with Visual and Textual Prompts
Authors: Shiyi Du, Xiaosong Wang, Yongyi Lu, Yuyin Zhou, Shaoting Zhang, Alan Yuille, Kang Li, Zongwei Zhou
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Abstract
Image synthesis approaches, e.g., generative adversarial networks, have been popular as a form of data augmentation in medical image analysis tasks. It is primarily beneficial to overcome the shortage of publicly accessible data and associated quality annotations. However, the current techniques often lack control over the detailed contents in generated images, e.g., the type of disease patterns, the location of lesions, and attributes of the diagnosis. In this work, we adapt the latest advance in the generative model, i.e., the diffusion model, with the added control flow using lesion-specific visual and textual prompts for generating dermatoscopic images. We further demonstrate the advantage of our diffusion model-based framework over the classical generation models in both the image quality and boosting the segmentation performance on skin lesions. It can achieve a 9% increase in the SSIM image quality measure and an over 5% increase in Dice coefficients over the prior arts.
Fast, Expressive SE$(n)$ Equivariant Networks through Weight-Sharing in Position-Orientation Space
Authors: Erik J Bekkers, Sharvaree Vadgama, Rob D Hesselink, Putri A van der Linden, David W Romero
Subjects: Machine Learning (cs.LG); Group Theory (math.GR)
Abstract
Based on the theory of homogeneous spaces we derive \textit{geometrically optimal edge attributes} to be used within the flexible message passing framework. We formalize the notion of weight sharing in convolutional networks as the sharing of message functions over point-pairs that should be treated equally. We define equivalence classes of point-pairs that are identical up to a transformation in the group and derive attributes that uniquely identify these classes. Weight sharing is then obtained by conditioning message functions on these attributes. As an application of the theory, we develop an efficient equivariant group convolutional network for processing 3D point clouds. The theory of homogeneous spaces tells us how to do group convolutions with feature maps over the homogeneous space of positions $\mathbb{R}^3$, position and orientations $\mathbb{R}^3 {\times} S^2$, and the group SE$(3)$ itself. Among these, $\mathbb{R}^3 {\times} S^2$ is an optimal choice due to the ability to represent directional information, which $\mathbb{R}^3$ methods cannot, and it significantly enhances computational efficiency compared to indexing features on the full SE$(3)$ group. We empirically support this claim by reaching state-of-the-art results -- in accuracy and speed -- on three different benchmarks: interatomic potential energy prediction, trajectory forecasting in N-body systems, and generating molecules via equivariant diffusion models.
T$^3$Bench: Benchmarking Current Progress in Text-to-3D Generation
Authors: Yuze He, Yushi Bai, Matthieu Lin, Wang Zhao, Yubin Hu, Jenny Sheng, Ran Yi, Juanzi Li, Yong-Jin Liu
Subjects: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL); Machine Learning (cs.LG)
Abstract
Recent methods in text-to-3D leverage powerful pretrained diffusion models to optimize NeRF. Notably, these methods are able to produce high-quality 3D scenes without training on 3D data. Due to the open-ended nature of the task, most studies evaluate their results with subjective case studies and user experiments, thereby presenting a challenge in quantitatively addressing the question: How has current progress in Text-to-3D gone so far? In this paper, we introduce T$^3$Bench, the first comprehensive text-to-3D benchmark containing diverse text prompts of three increasing complexity levels that are specially designed for 3D generation. To assess both the subjective quality and the text alignment, we propose two automatic metrics based on multi-view images produced by the 3D contents. The quality metric combines multi-view text-image scores and regional convolution to detect quality and view inconsistency. The alignment metric uses multi-view captioning and Large Language Model (LLM) evaluation to measure text-3D consistency. Both metrics closely correlate with different dimensions of human judgments, providing a paradigm for efficiently evaluating text-to-3D models. The benchmarking results, shown in Fig. 1, reveal performance differences among six prevalent text-to-3D methods. Our analysis further highlights the common struggles for current methods on generating surroundings and multi-object scenes, as well as the bottleneck of leveraging 2D guidance for 3D generation. Our project page is available at: https://t3bench.com.
Probing Intersectional Biases in Vision-Language Models with Counterfactual Examples
Abstract
While vision-language models (VLMs) have achieved remarkable performance improvements recently, there is growing evidence that these models also posses harmful biases with respect to social attributes such as gender and race. Prior studies have primarily focused on probing such bias attributes individually while ignoring biases associated with intersections between social attributes. This could be due to the difficulty of collecting an exhaustive set of image-text pairs for various combinations of social attributes from existing datasets. To address this challenge, we employ text-to-image diffusion models to produce counterfactual examples for probing intserctional social biases at scale. Our approach utilizes Stable Diffusion with cross attention control to produce sets of counterfactual image-text pairs that are highly similar in their depiction of a subject (e.g., a given occupation) while differing only in their depiction of intersectional social attributes (e.g., race & gender). We conduct extensive experiments using our generated dataset which reveal the intersectional social biases present in state-of-the-art VLMs.
Efficient-3DiM: Learning a Generalizable Single-image Novel-view Synthesizer in One Day
Authors: Yifan Jiang, Hao Tang, Jen-Hao Rick Chang, Liangchen Song, Zhangyang Wang, Liangliang Cao
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
The task of novel view synthesis aims to generate unseen perspectives of an object or scene from a limited set of input images. Nevertheless, synthesizing novel views from a single image still remains a significant challenge in the realm of computer vision. Previous approaches tackle this problem by adopting mesh prediction, multi-plain image construction, or more advanced techniques such as neural radiance fields. Recently, a pre-trained diffusion model that is specifically designed for 2D image synthesis has demonstrated its capability in producing photorealistic novel views, if sufficiently optimized on a 3D finetuning task. Although the fidelity and generalizability are greatly improved, training such a powerful diffusion model requires a vast volume of training data and model parameters, resulting in a notoriously long time and high computational costs. To tackle this issue, we propose Efficient-3DiM, a simple but effective framework to learn a single-image novel-view synthesizer. Motivated by our in-depth analysis of the inference process of diffusion models, we propose several pragmatic strategies to reduce the training overhead to a manageable scale, including a crafted timestep sampling strategy, a superior 3D feature extractor, and an enhanced training scheme. When combined, our framework is able to reduce the total training time from 10 days to less than 1 day, significantly accelerating the training process under the same computational platform (one instance with 8 Nvidia A100 GPUs). Comprehensive experiments are conducted to demonstrate the efficiency and generalizability of our proposed method.
Consistent-1-to-3: Consistent Image to 3D View Synthesis via Geometry-aware Diffusion Models
Authors: Jianglong Ye, Peng Wang, Kejie Li, Yichun Shi, Heng Wang
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Zero-shot novel view synthesis (NVS) from a single image is an essential problem in 3D object understanding. While recent approaches that leverage pre-trained generative models can synthesize high-quality novel views from in-the-wild inputs, they still struggle to maintain 3D consistency across different views. In this paper, we present Consistent-1-to-3, which is a generative framework that significantly mitigate this issue. Specifically, we decompose the NVS task into two stages: (i) transforming observed regions to a novel view, and (ii) hallucinating unseen regions. We design a scene representation transformer and view-conditioned diffusion model for performing these two stages respectively. Inside the models, to enforce 3D consistency, we propose to employ epipolor-guided attention to incorporate geometry constraints, and multi-view attention to better aggregate multi-view information. Finally, we design a hierarchy generation paradigm to generate long sequences of consistent views, allowing a full 360 observation of the provided object image. Qualitative and quantitative evaluation over multiple datasets demonstrate the effectiveness of the proposed mechanisms against state-of-the-art approaches. Our project page is at https://jianglongye.com/consistent123/
Keyword: adaptive
Shaping the Epochal Individuality and Generality: The Temporal Dynamics of Uncertainty and Prediction Error in Musical Improvisation
Authors: Tatsuya Daikoku
Subjects: Sound (cs.SD); Information Retrieval (cs.IR); Audio and Speech Processing (eess.AS)
Abstract
Musical improvisation, much like spontaneous speech, reveals intricate facets of the improviser's state of mind and emotional character. However, the specific musical components that reveal such individuality remain largely unexplored. Within the framework of brain's statistical learning and predictive processing, this study examined the temporal dynamics of uncertainty and surprise (prediction error) in a piece of musical improvisation. This study employed the HBSL model to analyze a corpus of 456 Jazz improvisations, spanning 1905 to 2009, from 78 distinct Jazz musicians. The results indicated distinctive temporal patterns of surprise and uncertainty, especially in pitch and pitch-rhythm sequences, revealing era-specific features from the early 20th to the 21st centuries. Conversely, rhythm sequences exhibited a consistent degree of uncertainty across eras. Further, the acoustic properties remain unchanged across different periods. These findings highlight the importance of how temporal dynamics of surprise and uncertainty in improvisational music change over periods, profoundly influencing the distinctive methodologies artists adopt for improvisation in each era. Further, it is suggested that the development of improvisational music can be attributed to the brain's adaptive statistical learning mechanisms, which constantly refine internal models to mirror the cultural and emotional nuances of their respective epochs. This study unravels the evolutionary trajectory of improvisational music and highlights the nuanced shifts artists employ to resonate with the cultural and emotional landscapes of their times.
Identifying Vulnerability Patches by Comprehending Code Commits with Comprehensive Change Contexts
Authors: Tianyu Chen, Lin Li, Taotao Qian, Zeyu Wang, Guangtai Liang, Ding Li, Qianxiang Wang, Tao Xie
Abstract
To help application developers apply vulnerability patches timely, security researchers maintain vulnerability databases such as National Vulnerability Database (NVD). By directly monitoring NVD with the name of each used library, application developers can be aware of vulnerabilities and their patches. Given that the monitoring results of vulnerability patches are unreliable due to patch incompleteness of NVD, existing approaches employ deep-learning (DL) models to identify additional vulnerability patches by determining whether a code commit fixes a vulnerability. However, these approaches suffer from low accuracy due to not considering code commits' comprehensive contexts such as control/data-flow contexts or method-invocation contexts. To improve accuracy, we design CompVPD, the first approach to identify vulnerability patches by fine-tuning a large language model (LLM) named StarCoder to comprehend code commits with comprehensive contexts. Considering that including comprehensive contexts needs to balance the context size and the training costs of LLM, CompVPD includes our two novel algorithms to generate comprehensive contexts within the given window size by removing irrelevant components (i.e., files, methods, and statements) and adaptively expanding each context. We empirically compare CompVPD with four state-of-the-art/practice (SOTA) approaches that identify vulnerability patches. The results show that CompVPD improves the AUC score by 11% and the F1 score by 30% when compared with the best scores of the SOTA approaches. Additionally, CompVPD provides high value to security practice by helping identify 20 vulnerability patches and 18 fixes of high-risk bugs from 2,500 recent code commits of five highly popular open-source projects.
SlowFormer: Universal Adversarial Patch for Attack on Compute and Energy Efficiency of Inference Efficient Vision Transformers
Abstract
Recently, there has been a lot of progress in reducing the computation of deep models at inference time. These methods can reduce both the computational needs and power usage of deep models. Some of these approaches adaptively scale the compute based on the input instance. We show that such models can be vulnerable to a universal adversarial patch attack, where the attacker optimizes for a patch that when pasted on any image, can increase the compute and power consumption of the model. We run experiments with three different efficient vision transformer methods showing that in some cases, the attacker can increase the computation to the maximum possible level by simply pasting a patch that occupies only 8\% of the image area. We also show that a standard adversarial training defense method can reduce some of the attack's success. We believe adaptive efficient methods will be necessary for the future to lower the power usage of deep models, so we hope our paper encourages the community to study the robustness of these methods and develop better defense methods for the proposed attack.
Generalization in diffusion models arises from geometry-adaptive harmonic representation
Authors: Zahra Kadkhodaie, Florentin Guth, Eero P. Simoncelli, Stéphane Mallat
Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Abstract
High-quality samples generated with score-based reverse diffusion algorithms provide evidence that deep neural networks (DNN) trained for denoising can learn high-dimensional densities, despite the curse of dimensionality. However, recent reports of memorization of the training set raise the question of whether these networks are learning the "true" continuous density of the data. Here, we show that two denoising DNNs trained on non-overlapping subsets of a dataset learn nearly the same score function, and thus the same density, with a surprisingly small number of training images. This strong generalization demonstrates an alignment of powerful inductive biases in the DNN architecture and/or training algorithm with properties of the data distribution. We analyze these, demonstrating that the denoiser performs a shrinkage operation in a basis adapted to the underlying image. Examination of these bases reveals oscillating harmonic structures along contours and in homogeneous image regions. We show that trained denoisers are inductively biased towards these geometry-adaptive harmonic representations by demonstrating that they arise even when the network is trained on image classes such as low-dimensional manifolds, for which the harmonic basis is suboptimal. Additionally, we show that the denoising performance of the networks is near-optimal when trained on regular image classes for which the optimal basis is known to be geometry-adaptive and harmonic.
AdaMerging: Adaptive Model Merging for Multi-Task Learning
Authors: Enneng Yang, Zhenyi Wang, Li Shen, Shiwei Liu, Guibing Guo, Xingwei Wang, Dacheng Tao
Subjects: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
Abstract
Multi-task learning (MTL) aims to empower a model to tackle multiple tasks simultaneously. A recent development known as task arithmetic has revealed that several models, each fine-tuned for distinct tasks, can be directly merged into a single model to execute MTL without necessitating a retraining process using the initial training data. Nevertheless, this direct addition of models often leads to a significant deterioration in the overall performance of the merged model. This decline occurs due to potential conflicts and intricate correlations among the multiple tasks. Consequently, the challenge emerges of how to merge pre-trained models more effectively without using their original training data. This paper introduces an innovative technique called Adaptive Model Merging (AdaMerging). This approach aims to autonomously learn the coefficients for model merging, either in a task-wise or layer-wise manner, without relying on the original training data. Specifically, our AdaMerging method operates as an automatic, unsupervised task arithmetic scheme. It leverages entropy minimization on unlabeled test samples from the multi-task setup as a surrogate objective function to iteratively refine the merging coefficients of the multiple models. Our experimental findings across eight tasks demonstrate the efficacy of the AdaMerging scheme we put forth. Compared to the current state-of-the-art task arithmetic merging scheme, AdaMerging showcases a remarkable 11\% improvement in performance. Notably, AdaMerging also exhibits superior generalization capabilities when applied to unseen downstream tasks. Furthermore, it displays a significantly enhanced robustness to data distribution shifts that may occur during the testing phase.
A priori error analysis of multirate time-stepping schemes for two-phase flow problems
Abstract
We present a priori error estimates for a multirate time-stepping scheme for coupled differential equations. The discretization is based on Galerkin methods in time using two different time meshes for two parts of the problem. We aim at surface coupled multiphysics problems like two-phase flows. Special focus is on the handling of the interface coupling to guarantee a coercive formulation as key to optimal order error estimates. In a sequence of increasing complexity, we begin with the coupling of two ordinary differential equations, coupled heat conduction equation, and finally a coupled Stokes problem. For this we show optimal multi-rate estimates in velocity and a suboptimal result in pressure. The a priori estimates prove that the multirate method decouples the two subproblems exactly. This is the basis for adaptive methods which can choose optimal lattices for the respective subproblems.
Adaptive Spatio-Temporal Voxels Based Trajectory Planning for Autonomous Driving in Highway Traffic Flow
Abstract
Trajectory planning is crucial for the safe driving of autonomous vehicles in highway traffic flow. Currently, some advanced trajectory planning methods utilize spatio-temporal voxels to construct feasible regions and then convert trajectory planning into optimization problem solving based on the feasible regions. However, these feasible region construction methods cannot adapt to the changes in dynamic environments, making them difficult to apply in complex traffic flow. In this paper, we propose a trajectory planning method based on adaptive spatio-temporal voxels which improves the construction of feasible regions and trajectory optimization while maintaining the quadratic programming form. The method can adjust feasible regions and trajectory planning according to real-time traffic flow and environmental changes, realizing vehicles to drive safely in complex traffic flow. The proposed method has been tested in both open-loop and closed-loop environments, and the test results show that our method outperforms the current planning methods.
A Study of Quantisation-aware Training on Time Series Transformer Models for Resource-constrained FPGAs
Authors: Tianheng Ling, Chao Qian, Lukas Einhaus, Gregor Schiele
Abstract
This study explores the quantisation-aware training (QAT) on time series Transformer models. We propose a novel adaptive quantisation scheme that dynamically selects between symmetric and asymmetric schemes during the QAT phase. Our approach demonstrates that matching the quantisation scheme to the real data distribution can reduce computational overhead while maintaining acceptable precision. Moreover, our approach is robust when applied to real-world data and mixed-precision quantisation, where most objects are quantised to 4 bits. Our findings inform model quantisation and deployment decisions while providing a foundation for advancing quantisation techniques.
MedPrompt: Cross-Modal Prompting for Multi-Task Medical Image Translation
Authors: Xuhang Chen, Chi-Man Pun, Shuqiang Wang
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Cross-modal medical image translation is an essential task for synthesizing missing modality data for clinical diagnosis. However, current learning-based techniques have limitations in capturing cross-modal and global features, restricting their suitability to specific pairs of modalities. This lack of versatility undermines their practical usefulness, particularly considering that the missing modality may vary for different cases. In this study, we present MedPrompt, a multi-task framework that efficiently translates different modalities. Specifically, we propose the Self-adaptive Prompt Block, which dynamically guides the translation network towards distinct modalities. Within this framework, we introduce the Prompt Extraction Block and the Prompt Fusion Block to efficiently encode the cross-modal prompt. To enhance the extraction of global features across diverse modalities, we incorporate the Transformer model. Extensive experimental results involving five datasets and four pairs of modalities demonstrate that our proposed model achieves state-of-the-art visual quality and exhibits excellent generalization capability.
Exploring Federated Optimization by Reducing Variance of Adaptive Unbiased Client Sampling
Abstract
Federated Learning (FL) systems usually sample a fraction of clients to conduct a training process. Notably, the variance of global estimates for updating the global model built on information from sampled clients is highly related to federated optimization quality. This paper explores a line of "free" adaptive client sampling techniques in federated optimization, where the server builds promising sampling probability and reliable global estimates without requiring additional local communication and computation. We capture a minor variant in the sampling procedure and improve the global estimation accordingly. Based on that, we propose a novel sampler called K-Vib, which solves an online convex optimization respecting client sampling in federated optimization. It achieves improved a linear speed up on regret bound $\tilde{\mathcal{O}}\big(N^{\frac{1}{3}}T^{\frac{2}{3}}/K^{\frac{4}{3}}\big)$ with communication budget $K$. As a result, it significantly improves the performance of federated optimization. Theoretical improvements and intensive experiments on classic federated tasks demonstrate our findings.
Tackling Hybrid Heterogeneity on Federated Optimization via Gradient Diversity Maximization
Abstract
Federated learning refers to a distributed machine learning paradigm in which data samples are decentralized and distributed among multiple clients. These samples may exhibit statistical heterogeneity, which refers to data distributions are not independent and identical across clients. Additionally, system heterogeneity, or variations in the computational power of the clients, introduces biases into federated learning. The combined effects of statistical and system heterogeneity can significantly reduce the efficiency of federated optimization. However, the impact of hybrid heterogeneity is not rigorously discussed. This paper explores how hybrid heterogeneity affects federated optimization by investigating server-side optimization. The theoretical results indicate that adaptively maximizing gradient diversity in server update direction can help mitigate the potential negative consequences of hybrid heterogeneity. To this end, we introduce a novel server-side gradient-based optimizer \textsc{FedAWARE} with theoretical guarantees provided. Intensive experiments in heterogeneous federated settings demonstrate that our proposed optimizer can significantly enhance the performance of federated learning across varying degrees of hybrid heterogeneity.
Understanding Pan-Sharpening via Generalized Inverse
Authors: Shiqi Liu, Yutong Bai, Xinyang Han, Alan Yuille
Subjects: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
Abstract
Pan-sharpening algorithm utilizes panchromatic image and multispectral image to obtain a high spatial and high spectral image. However, the optimizations of the algorithms are designed with different standards. We adopt the simple matrix equation to describe the Pan-sharpening problem. The solution existence condition and the acquirement of spectral and spatial resolution are discussed. A down-sampling enhancement method was introduced for better acquiring the spatial and spectral down-sample matrices. By the generalized inverse theory, we derived two forms of general inverse matrix formulations that can correspond to the two prominent classes of Pan-sharpening methods, that is, component substitution and multi-resolution analysis methods. Specifically, the Gram Schmidt Adaptive(GSA) was proved to follow the general inverse matrix formulation of component substitution. A model prior to the general inverse matrix of the spectral function was rendered. The theoretical errors are analyzed. Synthetic experiments and real data experiments are implemented. The proposed methods are better and sharper than other methods qualitatively in both synthetic and real experiments. The down-sample enhancement effect is shown of better results both quantitatively and qualitatively in real experiments. The generalized inverse matrix theory help us better understand the Pan-sharpening.
Dynamic Shuffle: An Efficient Channel Mixture Method
Abstract
The redundancy of Convolutional neural networks not only depends on weights but also depends on inputs. Shuffling is an efficient operation for mixing channel information but the shuffle order is usually pre-defined. To reduce the data-dependent redundancy, we devise a dynamic shuffle module to generate data-dependent permutation matrices for shuffling. Since the dimension of permutation matrix is proportional to the square of the number of input channels, to make the generation process efficiently, we divide the channels into groups and generate two shared small permutation matrices for each group, and utilize Kronecker product and cross group shuffle to obtain the final permutation matrices. To make the generation process learnable, based on theoretical analysis, softmax, orthogonal regularization, and binarization are employed to asymptotically approximate the permutation matrix. Dynamic shuffle adaptively mixes channel information with negligible extra computation and memory occupancy. Experiment results on image classification benchmark datasets CIFAR-10, CIFAR-100, Tiny ImageNet and ImageNet have shown that our method significantly increases ShuffleNets' performance. Adding dynamic generated matrix with learnable static matrix, we further propose static-dynamic-shuffle and show that it can serve as a lightweight replacement of ordinary pointwise convolution.
A Data-facilitated Numerical Method for Richards Equation to Model Water Flow Dynamics in Soil
Abstract
Root-zone soil moisture monitoring is essential for precision agriculture, smart irrigation, and drought prevention. Modeling the spatiotemporal water flow dynamics in soil is typically achieved by solving a hydrological model, such as the Richards equation which is a highly nonlinear partial differential equation (PDE). In this paper, we present a novel data-facilitated numerical method for solving the mixed-form Richards equation. This numerical method, which we call the D-GRW (Data-facilitated global Random Walk) method, synergistically integrates adaptive linearization scheme, neural networks, and global random walk in a finite volume discretization framework to produce accurate numerical solutions of the Richards equation with guaranteed convergence under reasonable assumptions. Through three illustrative examples, we demonstrate and discuss the superior accuracy and mass conservation performance of our D-GRW method and compare it with benchmark numerical methods and commercial solver.
A novel asymmetrical autoencoder with a sparsifying discrete cosine Stockwell transform layer for gearbox sensor data compression
Abstract
The lack of an efficient compression model remains a challenge for the wireless transmission of gearbox data in non-contact gear fault diagnosis problems. In this paper, we present a signal-adaptive asymmetrical autoencoder with a transform domain layer to compress sensor signals. First, a new discrete cosine Stockwell transform (DCST) layer is introduced to replace linear layers in a multi-layer autoencoder. A trainable filter is implemented in the DCST domain by utilizing the multiplication property of the convolution. A trainable hard-thresholding layer is applied to reduce redundant data in the DCST layer to make the feature map sparse. In comparison to the linear layer, the DCST layer reduces the number of trainable parameters and improves the accuracy of data reconstruction. Second, training the autoencoder with a sparsifying DCST layer only requires a small number of datasets. The proposed method is superior to other autoencoder-based methods on the University of Connecticut (UoC) and Southeast University (SEU) gearbox datasets, as the average quality score is improved by 2.00% at the lowest and 32.35% at the highest with a limited number of training samples
Immersive ExaBrick: Visualizing Large AMR Data in the CAVE
Authors: Zhaoyang Wang, Stefan Wesner, Stefan Zellmann
Abstract
Rendering large adaptive mesh refinement (AMR) data in real-time in virtual reality (VR) environments is a complex challenge that demands sophisticated techniques and tools. The proposed solution harnesses the ExaBrick framework and integrates it as a plugin in COVISE, a robust visualization system equipped with the VR-centric OpenCOVER render module. This setup enables direct navigation and interaction within the rendered volume in a VR environment. The user interface incorporates rendering options and functions, ensuring a smooth and interactive experience. We show that high-quality volume rendering of AMR data in VR environments at interactive rates is possible using GPUs.
Optimal Transport with Adaptive Regularisation
Authors: Hugues Van Assel, Titouan Vayer, Remi Flamary, Nicolas Courty
Abstract
Regularising the primal formulation of optimal transport (OT) with a strictly convex term leads to enhanced numerical complexity and a denser transport plan. Many formulations impose a global constraint on the transport plan, for instance by relying on entropic regularisation. As it is more expensive to diffuse mass for outlier points compared to central ones, this typically results in a significant imbalance in the way mass is spread across the points. This can be detrimental for some applications where a minimum of smoothing is required per point. To remedy this, we introduce OT with Adaptive RegularIsation (OTARI), a new formulation of OT that imposes constraints on the mass going in or/and out of each point. We then showcase the benefits of this approach for domain adaptation.
Adaptive Landmark Color for AUV Docking in Visually Dynamic Environments
Abstract
Autonomous Underwater Vehicles (AUVs) conduct missions underwater without the need for human intervention. A docking station (DS) can extend mission times of an AUV by providing a location for the AUV to recharge its batteries and receive updated mission information. Various methods for locating and tracking a DS exist, but most rely on expensive acoustic sensors, or are vision-based, which is significantly affected by water quality. In this \doctype, we present a vision-based method that utilizes adaptive color LED markers and dynamic color filtering to maximize landmark visibility in varying water conditions. Both AUV and DS utilize cameras to determine the water background color in order to calculate the desired marker color. No communication between AUV and DS is needed to determine marker color. Experiments conducted in a pool and lake show our method performs 10 times better than static color thresholding methods as background color varies. DS detection is possible at a range of 5 meters in clear water with minimal false positives.
SNR-Adaptive Ranging Waveform Design Based on Ziv-Zakai Bound Optimization
Authors: Yifeng Xiong, Fan Liu
Subjects: Information Theory (cs.IT); Signal Processing (eess.SP)
Abstract
Location-awareness is essential in various wireless applications. The capability of performing precise ranging is substantial in achieving high-accuracy localization. Due to the notorious ambiguity phenomenon, optimal ranging waveforms should be adaptive to the signal-to-noise ratio (SNR). In this letter, we propose to use the Ziv-Zakai bound (ZZB) as the ranging performance metric, as well as an associated waveform design algorithm having theoretical guarantee of achieving the optimal ZZB at a given SNR. Numerical results suggest that, in stark contrast to the well-known high-SNR design philosophy, the detection probability of the ranging signal becomes more important than the resolution in the low-SNR regime.
Towards Fully Adaptive Regret Minimization in Heavy-Tailed Bandits
Authors: Gianmarco Genalti, Lupo Marsigli, Nicola Gatti, Alberto Maria Metelli
Abstract
Heavy-tailed distributions naturally arise in many settings, from finance to telecommunications. While regret minimization under sub-Gaussian or bounded support rewards has been widely studied, learning on heavy-tailed distributions only gained popularity over the last decade. In the stochastic heavy-tailed bandit problem, an agent learns under the assumption that the distributions have finite moments of maximum order $1+\epsilon$ which are uniformly bounded by a constant $u$, for some $\epsilon \in (0,1]$. To the best of our knowledge, literature only provides algorithms requiring these two quantities as an input. In this paper, we study the stochastic adaptive heavy-tailed bandit, a variation of the standard setting where both $\epsilon$ and $u$ are unknown to the agent. We show that adaptivity comes at a cost, introducing two lower bounds on the regret of any adaptive algorithm, implying a higher regret w.r.t. the standard setting. Finally, we introduce a specific distributional assumption and provide Adaptive Robust UCB, a regret minimization strategy matching the known lower bound for the heavy-tailed MAB problem.
Understanding In-Context Learning in Transformers and LLMs by Learning to Learn Discrete Functions
Authors: Satwik Bhattamishra, Arkil Patel, Phil Blunsom, Varun Kanade
Subjects: Machine Learning (cs.LG); Computation and Language (cs.CL)
Abstract
In order to understand the in-context learning phenomenon, recent works have adopted a stylized experimental framework and demonstrated that Transformers can learn gradient-based learning algorithms for various classes of real-valued functions. However, the limitations of Transformers in implementing learning algorithms, and their ability to learn other forms of algorithms are not well understood. Additionally, the degree to which these capabilities are confined to attention-based models is unclear. Furthermore, it remains to be seen whether the insights derived from these stylized settings can be extrapolated to pretrained Large Language Models (LLMs). In this work, we take a step towards answering these questions by demonstrating the following: (a) On a test-bed with a variety of Boolean function classes, we find that Transformers can nearly match the optimal learning algorithm for 'simpler' tasks, while their performance deteriorates on more 'complex' tasks. Additionally, we find that certain attention-free models perform (almost) identically to Transformers on a range of tasks. (b) When provided a teaching sequence, i.e. a set of examples that uniquely identifies a function in a class, we show that Transformers learn more sample-efficiently. Interestingly, our results show that Transformers can learn to implement two distinct algorithms to solve a single task, and can adaptively select the more sample-efficient algorithm depending on the sequence of in-context examples. (c) Lastly, we show that extant LLMs, e.g. LLaMA-2, GPT-4, can compete with nearest-neighbor baselines on prediction tasks that are guaranteed to not be in their training set.
Keyword: quantization
Mixture of Quantized Experts (MoQE): Complementary Effect of Low-bit Quantization and Robustness
Authors: Young Jin Kim, Raffy Fahim, Hany Hassan Awadalla
Subjects: Machine Learning (cs.LG); Computation and Language (cs.CL)
Abstract
Large Mixture of Experts (MoE) models could achieve state-of-the-art quality on various language tasks, including machine translation task, thanks to the efficient model scaling capability with expert parallelism. However, it has brought a fundamental issue of larger memory consumption and increased memory bandwidth bottleneck at deployment time. In this paper, we propose Mixture of Quantized Experts (MoQE) which is a simple weight-only quantization method applying ultra low-bit down to 2-bit quantizations only to expert weights for mitigating the increased memory and latency issues of MoE models. We show that low-bit quantization together with the MoE architecture delivers a reliable model performance while reducing the memory size significantly even without any additional training in most cases. In particular, expert layers in MoE models are much more robust to the quantization than conventional feedforward networks (FFN) layers. In our comprehensive analysis, we show that MoE models with 2-bit expert weights can deliver better model performance than the dense model trained on the same dataset. As a result of low-bit quantization, we show the model size can be reduced by 79.6% of the original half precision floating point (fp16) MoE model. Combined with an optimized GPU runtime implementation, it also achieves 1.24X speed-up on A100 GPUs.
Spectral vs Energy Efficiency in 6G: Impact of the Receiver Front-End
Authors: Angel Lozano, Sundeep Rangan
Subjects: Information Theory (cs.IT); Signal Processing (eess.SP)
Abstract
This article puts the spotlight on the receiver front-end (RFE), an integral part of any wireless device that information theory typically idealizes into a mere addition of noise. While this idealization was sound in the past, as operating frequencies, bandwidths, and antenna counts rise, a soaring amount of power is required for the RFE to behave accordingly. Containing this surge in power expenditure exposes a harsher behavior on the part of the RFE (more noise, nonlinearities, and coarse quantization), setting up a tradeoff between the spectral efficiency under such nonidealities and the efficiency in the use of energy by the RFE. With the urge for radically better power consumptions and energy efficiencies in 6G, this emerges as an issue on which information theory can cast light at a fundamental level. More broadly, this article advocates the interest of having information theory embrace the device power consumption in its analyses. In turn, this calls for new models and abstractions such as the ones herein put together for the RFE, and for a more holistic perspective.
Foundation Reinforcement Learning: towards Embodied Generalist Agents with Foundation Prior Assistance
Authors: Weirui Ye, Yunsheng Zhang, Mengchen Wang, Shengjie Wang, Xianfan Gu, Pieter Abbeel, Yang Gao
Abstract
Recently, people have shown that large-scale pre-training from internet-scale data is the key to building generalist models, as witnessed in NLP. To build embodied generalist agents, we and many other researchers hypothesize that such foundation prior is also an indispensable component. However, it is unclear what is the proper concrete form to represent those embodied foundation priors and how they should be used in the downstream task. In this paper, we propose an intuitive and effective set of embodied priors that consist of foundation policy, value, and success reward. The proposed priors are based on the goal-conditioned MDP. To verify their effectiveness, we instantiate an actor-critic method assisted by the priors, called Foundation Actor-Critic (FAC). We name our framework as Foundation Reinforcement Learning (FRL), since it completely relies on embodied foundation priors to explore, learn and reinforce. The benefits of FRL are threefold. (1) Sample efficient. With foundation priors, FAC learns significantly faster than traditional RL. Our evaluation on the Meta-World has proved that FAC can achieve 100% success rates for 7/8 tasks under less than 200k frames, which outperforms the baseline method with careful manual-designed rewards under 1M frames. (2) Robust to noisy priors. Our method tolerates the unavoidable noise in embodied foundation models. We show that FAC works well even under heavy noise or quantization errors. (3) Minimal human intervention: FAC completely learns from the foundation priors, without the need of human-specified dense reward, or providing teleoperated demos. Thus, FAC can be easily scaled up. We believe our FRL framework could enable the future robot to autonomously explore and learn without human intervention in the physical world. In summary, our proposed FRL is a novel and powerful learning paradigm, towards achieving embodied generalist agents.
Soft Convex Quantization: Revisiting Vector Quantization with Convex Optimization
Abstract
Vector Quantization (VQ) is a well-known technique in deep learning for extracting informative discrete latent representations. VQ-embedded models have shown impressive results in a range of applications including image and speech generation. VQ operates as a parametric K-means algorithm that quantizes inputs using a single codebook vector in the forward pass. While powerful, this technique faces practical challenges including codebook collapse, non-differentiability and lossy compression. To mitigate the aforementioned issues, we propose Soft Convex Quantization (SCQ) as a direct substitute for VQ. SCQ works like a differentiable convex optimization (DCO) layer: in the forward pass, we solve for the optimal convex combination of codebook vectors that quantize the inputs. In the backward pass, we leverage differentiability through the optimality conditions of the forward solution. We then introduce a scalable relaxation of the SCQ optimization and demonstrate its efficacy on the CIFAR-10, GTSRB and LSUN datasets. We train powerful SCQ autoencoder models that significantly outperform matched VQ-based architectures, observing an order of magnitude better image reconstruction and codebook usage with comparable quantization runtime.
Keyword: efficient
MuSe-GNN: Learning Unified Gene Representation From Multimodal Biological Graph Data
Consistency Trajectory Models: Learning Probability Flow ODE Trajectory of Diffusion
Expert enhanced dynamic time warping based anomaly detection
A Comparison of Mesh-Free Differentiable Programming and Data-Driven Strategies for Optimal Control under PDE Constraints
Online Proactive Multi-Task Assignment with Resource Availability Anticipation
Event-Enhanced Multi-Modal Spiking Neural Network for Dynamic Obstacle Avoidance
Linear Recurrent Units for Sequential Recommendation
ScaleNet: An Unsupervised Representation Learning Method for Limited Information
Mixture of Quantized Experts (MoQE): Complementary Effect of Low-bit Quantization and Robustness
Machine learning assist nyc subway navigation safer and faster
Feather: An Elegant Solution to Effective DNN Sparsification
Impact of geography on the importance of parameters in infectious disease models
Distributionally Safe Reinforcement Learning under Model Uncertainty: A Single-Level Approach by Differentiable Convex Programming
Parameterized Model-checking of Discrete-Timed Networks and Symmetric-Broadcast Systems
Prompting-based Efficient Temporal Domain Generalization
Computationally Efficient Chance Constrained Covariance Control with Output Feedback
Ophiuchus: Scalable Modeling of Protein Structures through Hierarchical Coarse-graining SO(3)-Equivariant Autoencoders
Federated Conditional Stochastic Optimization
SlowFormer: Universal Adversarial Patch for Attack on Compute and Energy Efficiency of Inference Efficient Vision Transformers
Exact and soft boundary conditions in Physics-Informed Neural Networks for the Variable Coefficient Poisson equation
Convergence Analysis and Latency Minimization for Semi-Federated Learning in Massive IoT Networks
NOLA: Networks as Linear Combination of Low Rank Random Basis
Improving Drumming Robot Via Attention Transformer Network
ReForm-Eval: Evaluating Large Vision Language Models via Unified Re-Formulation of Task-Oriented Benchmarks
AdaMerging: Adaptive Model Merging for Multi-Task Learning
ViT-ReciproCAM: Gradient and Attention-Free Visual Explanations for Vision Transformer
Hybrid Inception Architecture with Residual Connection: Fine-tuned Inception-ResNet Deep Learning Model for Lung Inflammation Diagnosis from Chest Radiographs
Multi-Agent Reinforcement Learning for Power Grid Topology Optimization
Top-k contrast order-preserving pattern mining for time series classification
Multi-rules mining algorithm for combinatorially exploded decision trees with modified Aitchison-Aitken function-based Bayesian optimization
Foundation Reinforcement Learning: towards Embodied Generalist Agents with Foundation Prior Assistance
Deformation-Invariant Neural Network and Its Applications in Distorted Image Restoration and Analysis
Blend: A Unified Data Discovery System
MedPrompt: Cross-Modal Prompting for Multi-Task Medical Image Translation
Land-cover change detection using paired OpenStreetMap data and optical high-resolution imagery via object-guided Transformer
GETAvatar: Generative Textured Meshes for Animatable Human Avatars
Multi-resolution HuBERT: Multi-resolution Speech Self-Supervised Learning with Masked Unit Prediction
Leveraging Temporal Graph Networks Using Module Decoupling
Optimal Collaborative Transportation for Under-Capacitated Vehicle Routing Problems using Aerial Drone Swarms
Comparative Analysis of Imbalanced Malware Byteplot Image Classification using Transfer Learning
Likelihood-Based Methods Improve Parameter Estimation in Opinion Dynamics Models
Spike Accumulation Forwarding for Effective Training of Spiking Neural Networks
Dynamic Shuffle: An Efficient Channel Mixture Method
R-LGP: A Reachability-guided Logic-geometric Programming Framework for Optimal Task and Motion Planning on Mobile Manipulators
Everest: GPU-Accelerated System For Mining Temporal Motifs
Delving into CLIP latent space for Video Anomaly Recognition
A novel asymmetrical autoencoder with a sparsifying discrete cosine Stockwell transform layer for gearbox sensor data compression
Approximating Robot Configuration Spaces with few Convex Sets using Clique Covers of Visibility Graphs
A Grammatical Compositional Model for Video Action Detection
CoLiDE: Concomitant Linear DAG Estimation
Use Your INSTINCT: INSTruction optimization usIng Neural bandits Coupled with Transformers
Boosting Dermatoscopic Lesion Segmentation via Diffusion Models with Visual and Textual Prompts
Online Constraint Tightening in Stochastic Model Predictive Control: A Regression Approach
Proximal Policy Optimization-Based Reinforcement Learning Approach for DC-DC Boost Converter Control: A Comparative Evaluation Against Traditional Control Techniques
Fast, Expressive SE$(n)$ Equivariant Networks through Weight-Sharing in Position-Orientation Space
T$^3$Bench: Benchmarking Current Progress in Text-to-3D Generation
Never Train from Scratch: Fair Comparison of Long-Sequence Models Requires Data-Driven Priors
Exploring API Capabilities with Fieldwire
ECoFLaP: Efficient Coarse-to-Fine Layer-Wise Pruning for Vision-Language Models
From Words to Watts: Benchmarking the Energy Costs of Large Language Model Inference
Efficient-3DiM: Learning a Generalizable Single-image Novel-view Synthesizer in One Day
Understanding In-Context Learning in Transformers and LLMs by Learning to Learn Discrete Functions
Human-oriented Representation Learning for Robotic Manipulation
Keyword: faster
SE(3)-Stochastic Flow Matching for Protein Backbone Generation
Practical, Private Assurance of the Value of Collaboration
A Faster Deterministic Approximation Algorithm for TTP-2
On the Last-iterate Convergence in Time-varying Zero-sum Games: Extra Gradient Succeeds where Optimism Fails
Foundation Reinforcement Learning: towards Embodied Generalist Agents with Foundation Prior Assistance
ED-NeRF: Efficient Text-Guided Editing of 3D Scene using Latent Space NeRF
Stability Improvements for Fast Matrix Multiplication
FroSSL: Frobenius Norm Minimization for Self-Supervised Learning
SemiReward: A General Reward Model for Semi-supervised Learning
Retrieval meets Long Context Large Language Models
Keyword: mobile
Online Proactive Multi-Task Assignment with Resource Availability Anticipation
Event-Enhanced Multi-Modal Spiking Neural Network for Dynamic Obstacle Avoidance
Automatic Multi-Path Web Story Creation from a Structural Article
ShaSTA-Fuse: Camera-LiDAR Sensor Fusion to Model Shape and Spatio-Temporal Affinities for 3D Multi-Object Tracking
Hybrid Inception Architecture with Residual Connection: Fine-tuned Inception-ResNet Deep Learning Model for Lung Inflammation Diagnosis from Chest Radiographs
High order numerical methods based on quadratic spline collocation method and averaged L1 scheme for the variable-order time fractional mobile/immobile diffusion equation
R-LGP: A Reachability-guided Logic-geometric Programming Framework for Optimal Task and Motion Planning on Mobile Manipulators
Keyword: pruning
Junk DNA Hypothesis: A Task-Centric Angle of LLM Pre-trained Weights through Sparsity
Feather: An Elegant Solution to Effective DNN Sparsification
Convergence Analysis and Latency Minimization for Semi-Federated Learning in Massive IoT Networks
Top-k contrast order-preserving pattern mining for time series classification
Co-Optimizing Cache Partitioning and Multi-Core Task Scheduling: Exploit Cache Sensitivity or Not?
ECoFLaP: Efficient Coarse-to-Fine Layer-Wise Pruning for Vision-Language Models
Keyword: diffusion
Consistency Trajectory Models: Learning Probability Flow ODE Trajectory of Diffusion
Stochastic force inference via density estimation
SE(3)-Stochastic Flow Matching for Protein Backbone Generation
FT-Shield: A Watermark Against Unauthorized Fine-tuning in Text-to-Image Diffusion Models
EditVal: Benchmarking Diffusion Based Text-Guided Image Editing Methods
Learning to Reach Goals via Diffusion
Ophiuchus: Scalable Modeling of Protein Structures through Hierarchical Coarse-graining SO(3)-Equivariant Autoencoders
MedDiffusion: Boosting Health Risk Prediction via Diffusion-based Data Augmentation
Generalization in diffusion models arises from geometry-adaptive harmonic representation
SweetDreamer: Aligning Geometric Priors in 2D Diffusion for Consistent Text-to-3D
MagicDrive: Street View Generation with Diverse 3D Geometry Control
On Memorization in Diffusion Models
Diffusion Generative Flow Samplers: Improving learning signals through partial trajectory optimization
Robust Tracking for a 3D Diffusion Equation: Controlling Seismicity Rate in Geothermal Reservoirs
ED-NeRF: Efficient Text-Guided Editing of 3D Scene using Latent Space NeRF
High order numerical methods based on quadratic spline collocation method and averaged L1 scheme for the variable-order time fractional mobile/immobile diffusion equation
Magicremover: Tuning-free Text-guided Image inpainting with Diffusion Models
Boosting Dermatoscopic Lesion Segmentation via Diffusion Models with Visual and Textual Prompts
Fast, Expressive SE$(n)$ Equivariant Networks through Weight-Sharing in Position-Orientation Space
T$^3$Bench: Benchmarking Current Progress in Text-to-3D Generation
Probing Intersectional Biases in Vision-Language Models with Counterfactual Examples
Efficient-3DiM: Learning a Generalizable Single-image Novel-view Synthesizer in One Day
Consistent-1-to-3: Consistent Image to 3D View Synthesis via Geometry-aware Diffusion Models
Keyword: adaptive
Shaping the Epochal Individuality and Generality: The Temporal Dynamics of Uncertainty and Prediction Error in Musical Improvisation
Identifying Vulnerability Patches by Comprehending Code Commits with Comprehensive Change Contexts
SlowFormer: Universal Adversarial Patch for Attack on Compute and Energy Efficiency of Inference Efficient Vision Transformers
Generalization in diffusion models arises from geometry-adaptive harmonic representation
AdaMerging: Adaptive Model Merging for Multi-Task Learning
A priori error analysis of multirate time-stepping schemes for two-phase flow problems
Adaptive Spatio-Temporal Voxels Based Trajectory Planning for Autonomous Driving in Highway Traffic Flow
A Study of Quantisation-aware Training on Time Series Transformer Models for Resource-constrained FPGAs
MedPrompt: Cross-Modal Prompting for Multi-Task Medical Image Translation
Exploring Federated Optimization by Reducing Variance of Adaptive Unbiased Client Sampling
Tackling Hybrid Heterogeneity on Federated Optimization via Gradient Diversity Maximization
Understanding Pan-Sharpening via Generalized Inverse
Dynamic Shuffle: An Efficient Channel Mixture Method
A Data-facilitated Numerical Method for Richards Equation to Model Water Flow Dynamics in Soil
A novel asymmetrical autoencoder with a sparsifying discrete cosine Stockwell transform layer for gearbox sensor data compression
Immersive ExaBrick: Visualizing Large AMR Data in the CAVE
Optimal Transport with Adaptive Regularisation
Adaptive Landmark Color for AUV Docking in Visually Dynamic Environments
SNR-Adaptive Ranging Waveform Design Based on Ziv-Zakai Bound Optimization
Towards Fully Adaptive Regret Minimization in Heavy-Tailed Bandits
Understanding In-Context Learning in Transformers and LLMs by Learning to Learn Discrete Functions
Keyword: quantization
Mixture of Quantized Experts (MoQE): Complementary Effect of Low-bit Quantization and Robustness
Spectral vs Energy Efficiency in 6G: Impact of the Receiver Front-End
Foundation Reinforcement Learning: towards Embodied Generalist Agents with Foundation Prior Assistance
Soft Convex Quantization: Revisiting Vector Quantization with Convex Optimization