Abstract
Deep neural networks have gained great success due to the increasing amounts of data, and diverse effective neural network designs. However, it also brings a heavy computing burden as the amount of training data is proportional to the training time. In addition, a well-behaved model requires repeated trials of different structure designs and hyper-parameters, which may take a large amount of time even with state-of-the-art (SOTA) hyper-parameter optimization (HPO) algorithms and neural architecture search (NAS) algorithms. In this paper, we propose an Automatic Selection of Proxy dataset framework (ASP) aimed to dynamically find the informative proxy subsets of training data at each epoch, reducing the training data size as well as saving the AutoML processing time. We verify the effectiveness and generalization of ASP on CIFAR10, CIFAR100, ImageNet16-120, and ImageNet-1k, across various public model benchmarks. The experiment results show that ASP can obtain better results than other data selection methods at all selection ratios. ASP can also enable much more efficient AutoML processing with a speedup of 2x-20x while obtaining better architectures and better hyper-parameters compared to utilizing the entire dataset.
On the Temperature of Bayesian Graph Neural Networks for Conformal Prediction
Abstract
Accurate uncertainty quantification in graph neural networks (GNNs) is essential, especially in high-stakes domains where GNNs are frequently employed. Conformal prediction (CP) offers a promising framework for quantifying uncertainty by providing $\textit{valid}$ prediction sets for any black-box model. CP ensures formal probabilistic guarantees that a prediction set contains a true label with a desired probability. However, the size of prediction sets, known as $\textit{inefficiency}$, is influenced by the underlying model and data generating process. On the other hand, Bayesian learning also provides a credible region based on the estimated posterior distribution, but this region is $\textit{well-calibrated}$ only when the model is correctly specified. Building on a recent work that introduced a scaling parameter for constructing valid credible regions from posterior estimate, our study explores the advantages of incorporating a temperature parameter into Bayesian GNNs within CP framework. We empirically demonstrate the existence of temperatures that result in more efficient prediction sets. Furthermore, we conduct an analysis to identify the factors contributing to inefficiency and offer valuable insights into the relationship between CP performance and model calibration.
GenEval: An Object-Focused Framework for Evaluating Text-to-Image Alignment
Authors: Dhruba Ghosh, Hanna Hajishirzi, Ludwig Schmidt
Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Abstract
Recent breakthroughs in diffusion models, multimodal pretraining, and efficient finetuning have led to an explosion of text-to-image generative models. Given human evaluation is expensive and difficult to scale, automated methods are critical for evaluating the increasingly large number of new models. However, most current automated evaluation metrics like FID or CLIPScore only offer a holistic measure of image quality or image-text alignment, and are unsuited for fine-grained or instance-level analysis. In this paper, we introduce GenEval, an object-focused framework to evaluate compositional image properties such as object co-occurrence, position, count, and color. We show that current object detection models can be leveraged to evaluate text-to-image models on a variety of generation tasks with strong human agreement, and that other discriminative vision models can be linked to this pipeline to further verify properties like object color. We then evaluate several open-source text-to-image models and analyze their relative generative capabilities on our benchmark. We find that recent models demonstrate significant improvement on these tasks, though they are still lacking in complex capabilities such as spatial relations and attribute binding. Finally, we demonstrate how GenEval might be used to help discover existing failure modes, in order to inform development of the next generation of text-to-image models. Our code to run the GenEval framework is publicly available at https://github.com/djghosh13/geneval.
Value-Biased Maximum Likelihood Estimation for Model-based Reinforcement Learning in Discounted Linear MDPs
Authors: Yu-Heng Hung, Ping-Chun Hsieh, Akshay Mete, P. R. Kumar
Abstract
We consider the infinite-horizon linear Markov Decision Processes (MDPs), where the transition probabilities of the dynamic model can be linearly parameterized with the help of a predefined low-dimensional feature mapping. While the existing regression-based approaches have been theoretically shown to achieve nearly-optimal regret, they are computationally rather inefficient due to the need for a large number of optimization runs in each time step, especially when the state and action spaces are large. To address this issue, we propose to solve linear MDPs through the lens of Value-Biased Maximum Likelihood Estimation (VBMLE), which is a classic model-based exploration principle in the adaptive control literature for resolving the well-known closed-loop identification problem of Maximum Likelihood Estimation. We formally show that (i) VBMLE enjoys $\widetilde{O}(d\sqrt{T})$ regret, where $T$ is the time horizon and $d$ is the dimension of the model parameter, and (ii) VBMLE is computationally more efficient as it only requires solving one optimization problem in each time step. In our regret analysis, we offer a generic convergence result of MLE in linear MDPs through a novel supermartingale construct and uncover an interesting connection between linear MDPs and online learning, which could be of independent interest. Finally, the simulation results show that VBMLE significantly outperforms the benchmark method in terms of both empirical regret and computation time.
Efficient Online Learning with Offline Datasets for Infinite Horizon MDPs: A Bayesian Approach
Abstract
In this paper, we study the problem of efficient online reinforcement learning in the infinite horizon setting when there is an offline dataset to start with. We assume that the offline dataset is generated by an expert but with unknown level of competence, i.e., it is not perfect and not necessarily using the optimal policy. We show that if the learning agent models the behavioral policy (parameterized by a competence parameter) used by the expert, it can do substantially better in terms of minimizing cumulative regret, than if it doesn't do that. We establish an upper bound on regret of the exact informed PSRL algorithm that scales as $\tilde{O}(\sqrt{T})$. This requires a novel prior-dependent regret analysis of Bayesian online learning algorithms for the infinite horizon setting. We then propose an approximate Informed RLSVI algorithm that we can interpret as performing imitation learning with the offline dataset, and then performing online learning.
HMN: Generalization of Heterogeneous and Multi-layered Network
Abstract
A network may have different types of entities and their relations. Further, there could be additional layers of ties. The former is referred to as Heterogeneous networks, while the latter is known as Multi-layer networks. The present paper provides a generalized network model, namely, a Heterogeneous Multi-layered Network (HMN), which can simultaneously be multi-layered and heterogeneous. The model can represent homogeneous networks as well. We define different structural measures in an HMN. We proved that the sets of all homogeneous, heterogeneous and multi-layered networks are subsets of the set of all HMNs. Accordingly, we established the equivalency of the proposed structural measures of HMNs with that of homogeneous, heterogeneous, and multi-layered networks. Following that, we show how our proposed HMN is more efficient in tasks such as link prediction. In addition, we present a novel parameterized algorithm (with complexity analysis) for generating synthetic HMNs. The networks generated from our proposed algorithm are more consistent in modelling the layer-wise degree distribution of a real-world Twitter network (represented as HMN) than those generated by existing models. Moreover, we also show that our algorithm is more effective in modelling an air-transportation multiplex network when compared to an algorithm designed specifically for the task.
Towards Optimal Regret in Adversarial Linear MDPs with Bandit Feedback
Abstract
We study online reinforcement learning in linear Markov decision processes with adversarial losses and bandit feedback, without prior knowledge on transitions or access to simulators. We introduce two algorithms that achieve improved regret performance compared to existing approaches. The first algorithm, although computationally inefficient, ensures a regret of $\widetilde{\mathcal{O}}\left(\sqrt{K}\right)$, where $K$ is the number of episodes. This is the first result with the optimal $K$ dependence in the considered setting. The second algorithm, which is based on the policy optimization framework, guarantees a regret of $\widetilde{\mathcal{O}}\left(K^{\frac{3}{4}} \right)$ and is computationally efficient. Both our results significantly improve over the state-of-the-art: a computationally inefficient algorithm by Kong et al. [2023] with $\widetilde{\mathcal{O}}\left(K^{\frac{4}{5}}+poly\left(\frac{1}{\lambda{\min}}\right) \right)$ regret, for some problem-dependent constant $\lambda{\min}$ that can be arbitrarily close to zero, and a computationally efficient algorithm by Sherman et al. [2023b] with $\widetilde{\mathcal{O}}\left(K^{\frac{6}{7}} \right)$ regret.
Personalized Soups: Personalized Large Language Model Alignment via Post-hoc Parameter Merging
Authors: Joel Jang, Seungone Kim, Bill Yuchen Lin, Yizhong Wang, Jack Hessel, Luke Zettlemoyer, Hannaneh Hajishirzi, Yejin Choi, Prithviraj Ammanabrolu
Abstract
While Reinforcement Learning from Human Feedback (RLHF) aligns Large Language Models (LLMs) with general, aggregate human preferences, it is suboptimal for learning diverse, individual perspectives. In this work, we study Reinforcement Learning from Personalized Human Feedback (RLPHF) problem, wherein LLMs are aligned to multiple (sometimes conflicting) preferences by modeling alignment as a Multi-Objective Reinforcement Learning (MORL) problem. Compared to strong single-objective baselines, we show that we can achieve personalized alignment by decomposing preferences into multiple dimensions. These dimensions are defined based on personalizations that are declared as desirable by the user. In this work, we show that they can be efficiently trained independently in a distributed manner and combined effectively post-hoc through parameter merging. The code is available at https://github.com/joeljang/RLPHF.
FreSh: A Lock-Free Data Series Index
Authors: Panagiota Fatourou, Eleftherios Kosmas, Themis Palpanas, George Paterakis
Subjects: Databases (cs.DB); Distributed, Parallel, and Cluster Computing (cs.DC)
Abstract
We present FreSh, a lock-free data series index that exhibits good performance (while being robust). FreSh is based on Refresh, which is a generic approach we have developed for supporting lock-freedom in an efficient way on top of any localityaware data series index. We believe Refresh is of independent interest and can be used to get well-performed lock-free versions of other locality-aware blocking data structures. For developing FreSh, we first studied in depth the design decisions of current state-of-the-art data series indexes, and the principles governing their performance. This led to a theoretical framework, which enables the development and analysis of data series indexes in a modular way. The framework allowed us to apply Refresh, repeatedly, to get lock-free versions of the different phases of a family of data series indexes. Experiments with several synthetic and real datasets illustrate that FreSh achieves performance that is as good as that of the state-of-the-art blocking in-memory data series index. This shows that the helping mechanisms of FreSh are light-weight, respecting certain principles that are crucial for performance in locality-aware data structures.This paper was published in SRDS 2023.
In defense of parameter sharing for model-compression
Abstract
When considering a model architecture, there are several ways to reduce its memory footprint. Historically, popular approaches included selecting smaller architectures and creating sparse networks through pruning. More recently, randomized parameter-sharing (RPS) methods have gained traction for model compression at start of training. In this paper, we comprehensively assess the trade-off between memory and accuracy across RPS, pruning techniques, and building smaller models. Our findings demonstrate that RPS, which is both data and model-agnostic, consistently outperforms/matches smaller models and all moderately informed pruning strategies, such as MAG, SNIP, SYNFLOW, and GRASP, across the entire compression range. This advantage becomes particularly pronounced in higher compression scenarios. Notably, even when compared to highly informed pruning techniques like Lottery Ticket Rewinding (LTR), RPS exhibits superior performance in high compression settings. This points out inherent capacity advantage that RPS enjoys over sparse models. Theoretically, we establish RPS as a superior technique in terms of memory-efficient representation when compared to pruning for linear models. This paper argues in favor of paradigm shift towards RPS based models. During our rigorous evaluation of RPS, we identified issues in the state-of-the-art RPS technique ROAST, specifically regarding stability (ROAST's sensitivity to initialization hyperparameters, often leading to divergence) and Pareto-continuity (ROAST's inability to recover the accuracy of the original model at zero compression). We provably address both of these issues. We refer to the modified RPS, which incorporates our improvements, as STABLE-RPS.
VKIE: The Application of Key Information Extraction on Video Text
Authors: Siyu An, Ye Liu, Haoyuan Peng, Di Yin
Subjects: Information Retrieval (cs.IR); Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM)
Abstract
Extracting structured information from videos is critical for numerous downstream applications in the industry. In this paper, we define a significant task of extracting hierarchical key information from visual texts on videos. To fulfill this task, we decouples it into four subtasks and introduce two implementation solutions called PipVKIE and UniVKIE. PipVKIE sequentially completes the four subtasks in continuous stages, while UniVKIE is improved by unifying all the subtasks into one backbone. Both PipVKIE and UniVKIE leverage multimodal information from vision, text, and coordinates for feature representation. Extensive experiments on one well-defined dataset demonstrate that our solutions can achieve remarkable performance and efficient inference speed. The code and dataset will be publicly available.
US Microelectronics Packaging Ecosystem: Challenges and Opportunities
Authors: Rouhan Noor, Himanandhan Reddy Kottur, Patrick J Craig, Liton Kumar Biswas, M Shafkat M Khan, Nitin Varshney, Hamed Dalir, Elif Akçalı, Bahar Motlagh, Charles Woychik, Yong-Kyu Yoon, Navid Asadizanjani
Subjects: Systems and Control (eess.SY); Cryptography and Security (cs.CR)
Abstract
The semiconductor industry is experiencing a significant shift from traditional methods of shrinking devices and reducing costs. Chip designers actively seek new technological solutions to enhance cost-effectiveness while incorporating more features into the silicon footprint. One promising approach is Heterogeneous Integration (HI), which involves advanced packaging techniques to integrate independently designed and manufactured components using the most suitable process technology. However, adopting HI introduces design and security challenges. To enable HI, research and development of advanced packaging is crucial. The existing research raises the possible security threats in the advanced packaging supply chain, as most of the Outsourced Semiconductor Assembly and Test (OSAT) facilities/vendors are offshore. To deal with the increasing demand for semiconductors and to ensure a secure semiconductor supply chain, there are sizable efforts from the United States (US) government to bring semiconductor fabrication facilities onshore. However, the US-based advanced packaging capabilities must also be ramped up to fully realize the vision of establishing a secure, efficient, resilient semiconductor supply chain. Our effort was motivated to identify the possible bottlenecks and weak links in the advanced packaging supply chain based in the US.
High Efficiency Polymer based Direct Multi-jet Impingement Cooling Solution for High Power Devices
Abstract
Liquid jet impingement cooling is an efficient cooling technique where the liquid coolant is directly ejected from nozzles on the chip backside resulting in a high cooling efficiency due to the absence of the TIM and the lateral temperature gradient. In literature, several Si-fabrication based impingement coolers with nozzle diameters of a few distributed returns or combination of micro-channels and impingement nozzles. The drawback of this Si processing of the cooler is the high fabrication cost. Other fabrication methods for nozzle diameters for ceramic and metal. Low cost fabrication methods, including injection molding and 3D printing have been introduced for much larger nozzle diameters (mm range) with larger cooler dimensions. These dimensions and processes are however not compatible with the chip packaging process flow. This PhD focuses on the modeling, design, fabrication and characterization of a micro-scale liquid impingement cooler using advanced, yet cost efficient, fabrication techniques. The main objectives are: (a) development of a modeling methodology to optimize the cooler geometry; (b) exploring low cost fabrication methods for the package level impingement jet cooler; (c) experimental thermal and hydraulic characterization and analysis of the fabricated coolers; (d) applying the direct impingement jet cooling solutions to different applications.
Prototype-based HyperAdapter for Sample-Efficient Multi-task Tuning
Authors: Hao Zhao, Jie Fu, Zhaofeng He
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
Abstract
Parameter-efficient fine-tuning (PEFT) has shown its effectiveness in adapting the pre-trained language models to downstream tasks while only updating a small number of parameters. Despite the success, most existing methods independently adapt to each task without considering knowledge transfer between tasks and are limited to low-data regimes. To overcome this issue, we propose Prototype-based HyperAdapter (PHA), a novel framework built on the adapter-tuning and hypernetwork. It introduces an instance-dense retriever and a prototypical hypernetwork to generate the conditional modules in a sample-efficient manner. This leads to comparable performance improvements against existing PEFT methods on multi-task learning and few-shot transfer learning. More importantly, when the available data size gets smaller, our method outperforms other strong baselines by a large margin. Based on our extensive empirical experiments across various datasets, we demonstrate that PHA strikes a better trade-off between trainable parameters, accuracy on stream tasks, and sample efficiency.
MixEdit: Revisiting Data Augmentation and Beyond for Grammatical Error Correction
Authors: Jingheng Ye, Yinghui Li, Yangning Li, Hai-Tao Zheng
Abstract
Data Augmentation through generating pseudo data has been proven effective in mitigating the challenge of data scarcity in the field of Grammatical Error Correction (GEC). Various augmentation strategies have been widely explored, most of which are motivated by two heuristics, i.e., increasing the distribution similarity and diversity of pseudo data. However, the underlying mechanism responsible for the effectiveness of these strategies remains poorly understood. In this paper, we aim to clarify how data augmentation improves GEC models. To this end, we introduce two interpretable and computationally efficient measures: Affinity and Diversity. Our findings indicate that an excellent GEC data augmentation strategy characterized by high Affinity and appropriate Diversity can better improve the performance of GEC models. Based on this observation, we propose MixEdit, a data augmentation approach that strategically and dynamically augments realistic data, without requiring extra monolingual corpora. To verify the correctness of our findings and the effectiveness of the proposed MixEdit, we conduct experiments on mainstream English and Chinese GEC datasets. The results show that MixEdit substantially improves GEC models and is complementary to traditional data augmentation methods.
PREM: A Simple Yet Effective Approach for Node-Level Graph Anomaly Detection
Authors: Junjun Pan, Yixin Liu, Yizhen Zheng, Shirui Pan
Abstract
Node-level graph anomaly detection (GAD) plays a critical role in identifying anomalous nodes from graph-structured data in various domains such as medicine, social networks, and e-commerce. However, challenges have arisen due to the diversity of anomalies and the dearth of labeled data. Existing methodologies - reconstruction-based and contrastive learning - while effective, often suffer from efficiency issues, stemming from their complex objectives and elaborate modules. To improve the efficiency of GAD, we introduce a simple method termed PREprocessing and Matching (PREM for short). Our approach streamlines GAD, reducing time and memory consumption while maintaining powerful anomaly detection capabilities. Comprising two modules - a pre-processing module and an ego-neighbor matching module - PREM eliminates the necessity for message-passing propagation during training, and employs a simple contrastive loss, leading to considerable reductions in training time and memory usage. Moreover, through rigorous evaluations of five real-world datasets, our method demonstrated robustness and effectiveness. Notably, when validated on the ACM dataset, PREM achieved a 5% improvement in AUC, a 9-fold increase in training speed, and sharply reduce memory usage compared to the most efficient baseline.
Improved Sample Complexity Analysis of Natural Policy Gradient Algorithm with General Parameterization for Infinite Horizon Discounted Reward Markov Decision Processes
Abstract
We consider the problem of designing sample efficient learning algorithms for infinite horizon discounted reward Markov Decision Process. Specifically, we propose the Accelerated Natural Policy Gradient (ANPG) algorithm that utilizes an accelerated stochastic gradient descent process to obtain the natural policy gradient. ANPG achieves $\mathcal{O}({\epsilon^{-2}})$ sample complexity and $\mathcal{O}(\epsilon^{-1})$ iteration complexity with general parameterization where $\epsilon$ defines the optimality error. This improves the state-of-the-art sample complexity by a $\log(\frac{1}{\epsilon})$ factor. ANPG is a first-order algorithm and unlike some existing literature, does not require the unverifiable assumption that the variance of importance sampling (IS) weights is upper bounded. In the class of Hessian-free and IS-free algorithms, ANPG beats the best-known sample complexity by a factor of $\mathcal{O}(\epsilon^{-\frac{1}{2}})$ and simultaneously matches their state-of-the-art iteration complexity.
Descriptive Knowledge Graph in Biomedical Domain
Authors: Kerui Zhu, Jie Huang, Kevin Chen-Chuan Chang
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
Abstract
We present a novel system that automatically extracts and generates informative and descriptive sentences from the biomedical corpus and facilitates the efficient search for relational knowledge. Unlike previous search engines or exploration systems that retrieve unconnected passages, our system organizes descriptive sentences as a relational graph, enabling researchers to explore closely related biomedical entities (e.g., diseases treated by a chemical) or indirectly connected entities (e.g., potential drugs for treating a disease). Our system also uses ChatGPT and a fine-tuned relation synthesis model to generate concise and reliable descriptive sentences from retrieved information, reducing the need for extensive human reading effort. With our system, researchers can easily obtain both high-level knowledge and detailed references and interactively steer to the information of interest. We spotlight the application of our system in COVID-19 research, illustrating its utility in areas such as drug repurposing and literature curation.
Quantum Acceleration of Infinite Horizon Average-Reward Reinforcement Learning
Abstract
This paper investigates the potential of quantum acceleration in addressing infinite horizon Markov Decision Processes (MDPs) to enhance average reward outcomes. We introduce an innovative quantum framework for the agent's engagement with an unknown MDP, extending the conventional interaction paradigm. Our approach involves the design of an optimism-driven tabular Reinforcement Learning algorithm that harnesses quantum signals acquired by the agent through efficient quantum mean estimation techniques. Through thorough theoretical analysis, we demonstrate that the quantum advantage in mean estimation leads to exponential advancements in regret guarantees for infinite horizon Reinforcement Learning. Specifically, the proposed Quantum algorithm achieves a regret bound of $\tilde{\mathcal{O}}(1)$, a significant improvement over the $\tilde{\mathcal{O}}(\sqrt{T})$ bound exhibited by classical counterparts.
Superiority of Softmax: Unveiling the Performance Edge Over Linear Attention
Authors: Yichuan Deng, Zhao Song, Tianyi Zhou
Subjects: Computation and Language (cs.CL); Machine Learning (cs.LG)
Abstract
Large transformer models have achieved state-of-the-art results in numerous natural language processing tasks. Among the pivotal components of the transformer architecture, the attention mechanism plays a crucial role in capturing token interactions within sequences through the utilization of softmax function. Conversely, linear attention presents a more computationally efficient alternative by approximating the softmax operation with linear complexity. However, it exhibits substantial performance degradation when compared to the traditional softmax attention mechanism. In this paper, we bridge the gap in our theoretical understanding of the reasons behind the practical performance gap between softmax and linear attention. By conducting a comprehensive comparative analysis of these two attention mechanisms, we shed light on the underlying reasons for why softmax attention outperforms linear attention in most scenarios.
Adaptation with Self-Evaluation to Improve Selective Prediction in LLMs
Authors: Jiefeng Chen, Jinsung Yoon, Sayna Ebrahimi, Sercan O Arik, Tomas Pfister, Somesh Jha
Subjects: Computation and Language (cs.CL); Machine Learning (cs.LG)
Abstract
Large language models (LLMs) have recently shown great advances in a variety of tasks, including natural language understanding and generation. However, their use in high-stakes decision-making scenarios is still limited due to the potential for errors. Selective prediction is a technique that can be used to improve the reliability of the LLMs by allowing them to abstain from making predictions when they are unsure of the answer. In this work, we propose a novel framework for adaptation with self-evaluation to improve the selective prediction performance of LLMs. Our framework is based on the idea of using parameter-efficient tuning to adapt the LLM to the specific task at hand while improving its ability to perform self-evaluation. We evaluate our method on a variety of question-answering (QA) datasets and show that it outperforms state-of-the-art selective prediction methods. For example, on the CoQA benchmark, our method improves the AUACC from 91.23% to 92.63% and improves the AUROC from 74.61% to 80.25%.
Runner re-identification from single-view video in the open-world setting
Abstract
In many sports, player re-identification is crucial for automatic video processing and analysis. However, most of the current studies on player re-identification in multi- or single-view sports videos focus on re-identification in the closed-world setting using labeled image dataset, and player re-identification in the open-world setting for automatic video analysis is not well developed. In this paper, we propose a runner re-identification system that directly processes single-view video to address the open-world setting. In the open-world setting, we cannot use labeled dataset and have to process video directly. The proposed system automatically processes raw video as input to identify runners, and it can identify runners even when they are framed out multiple times. For the automatic processing, we first detect the runners in the video using the pre-trained YOLOv8 and the fine-tuned EfficientNet. We then track the runners using ByteTrack and detect their shoes with the fine-tuned YOLOv8. Finally, we extract the image features of the runners using an unsupervised method using the gated recurrent unit autoencoder model. To improve the accuracy of runner re-identification, we use dynamic features of running sequence images. We evaluated the system on a running practice video dataset and showed that the proposed method identified runners with higher accuracy than one of the state-of-the-art models in unsupervised re-identification. We also showed that our unsupervised running dynamic feature extractor was effective for runner re-identification. Our runner re-identification system can be useful for the automatic analysis of running videos.
DPF-Nutrition: Food Nutrition Estimation via Depth Prediction and Fusion
Abstract
A reasonable and balanced diet is essential for maintaining good health. With the advancements in deep learning, automated nutrition estimation method based on food images offers a promising solution for monitoring daily nutritional intake and promoting dietary health. While monocular image-based nutrition estimation is convenient, efficient, and economical, the challenge of limited accuracy remains a significant concern. To tackle this issue, we proposed DPF-Nutrition, an end-to-end nutrition estimation method using monocular images. In DPF-Nutrition, we introduced a depth prediction module to generate depth maps, thereby improving the accuracy of food portion estimation. Additionally, we designed an RGB-D fusion module that combined monocular images with the predicted depth information, resulting in better performance for nutrition estimation. To the best of our knowledge, this was the pioneering effort that integrated depth prediction and RGB-D fusion techniques in food nutrition estimation. Comprehensive experiments performed on Nutrition5k evaluated the effectiveness and efficiency of DPF-Nutrition.
Quantify Health-Related Atomic Knowledge in Chinese Medical Large Language Models: A Computational Analysis
Authors: Yaxin Fan, Feng Jiang, Peifeng Li, Haizhou Li
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
Abstract
Large Language Models (LLMs) have the potential to revolutionize the way users self-diagnose through search engines by offering direct and efficient suggestions. Recent studies primarily focused on the quality of LLMs evaluated by GPT-4 or their ability to pass medical exams, no studies have quantified the extent of health-related atomic knowledge stored in LLMs' memory, which is the basis of LLMs to provide more factual suggestions. In this paper, we first constructed a benchmark, including the most common types of atomic knowledge in user self-diagnosis queries, with 17 atomic types and a total of 14, 048 pieces of atomic knowledge. Then, we evaluated both generic and specialized LLMs on the benchmark. The experimental results showcased that generic LLMs perform better than specialized LLMs in terms of atomic knowledge and instruction-following ability. Error analysis revealed that both generic and specialized LLMs are sycophantic, e.g., always catering to users' claims when it comes to unknown knowledge. Besides, generic LLMs showed stronger safety, which can be learned by specialized LLMs through distilled data. We further explored different types of data commonly adopted for fine-tuning specialized LLMs, i.e., real-world, semi-distilled, and distilled data, and found that distilled data can benefit LLMs most.
VST++: Efficient and Stronger Visual Saliency Transformer
Authors: Nian Liu, Ziyang Luo, Ni Zhang, Junwei Han
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
While previous CNN-based models have exhibited promising results for salient object detection (SOD), their ability to explore global long-range dependencies is restricted. Our previous work, the Visual Saliency Transformer (VST), addressed this constraint from a transformer-based sequence-to-sequence perspective, to unify RGB and RGB-D SOD. In VST, we developed a multi-task transformer decoder that concurrently predicts saliency and boundary outcomes in a pure transformer architecture. Moreover, we introduced a novel token upsampling method called reverse T2T for predicting a high-resolution saliency map effortlessly within transformer-based structures. Building upon the VST model, we further propose an efficient and stronger VST version in this work, i.e. VST++. To mitigate the computational costs of the VST model, we propose a Select-Integrate Attention (SIA) module, partitioning foreground into fine-grained segments and aggregating background information into a single coarse-grained token. To incorporate 3D depth information with low cost, we design a novel depth position encoding method tailored for depth maps. Furthermore, we introduce a token-supervised prediction loss to provide straightforward guidance for the task-related tokens. We evaluate our VST++ model across various transformer-based backbones on RGB, RGB-D, and RGB-T SOD benchmark datasets. Experimental results show that our model outperforms existing methods while achieving a 25% reduction in computational costs without significant performance compromise. The demonstrated strong ability for generalization, enhanced performance, and heightened efficiency of our VST++ model highlight its potential.
EchoScan: Scanning Complex Indoor Geometries via Acoustic Echoes
Abstract
Accurate estimation of indoor space geometries is vital for constructing precise digital twins, whose broad industrial applications include navigation in unfamiliar environments and efficient evacuation planning, particularly in low-light conditions. This study introduces EchoScan, a deep neural network model that utilizes acoustic echoes to perform room geometry inference. Conventional sound-based techniques rely on estimating geometry-related room parameters such as wall position and room size, thereby limiting the diversity of inferable room geometries. Contrarily, EchoScan overcomes this limitation by directly inferring room floorplans and heights, thereby enabling it to handle rooms with arbitrary shapes, including curved walls. The key innovation of EchoScan is its ability to analyze the complex relationship between low- and high-order reflections in room impulse responses (RIRs) using a multi-aggregation module. The analysis of high-order reflections also enables it to infer complex room shapes when echoes are unobservable from the position of an audio device. Herein, EchoScan was trained and evaluated using RIRs synthesized from complex environments, including the Manhattan and Atlanta layouts, employing a practical audio device configuration compatible with commercial, off-the-shelf devices. Compared with vision-based methods, EchoScan demonstrated outstanding geometry estimation performance in rooms with various shapes.
Investigating Uncertainty Calibration of Aligned Language Models under the Multiple-Choice Setting
Abstract
Despite the significant progress made in practical applications of aligned language models (LMs), they tend to be overconfident in output answers compared to the corresponding pre-trained LMs. In this work, we systematically evaluate the impact of the alignment process on logit-based uncertainty calibration of LMs under the multiple-choice setting. We first conduct a thoughtful empirical study on how aligned LMs differ in calibration from their pre-trained counterparts. Experimental results reveal that there are two distinct uncertainties in LMs under the multiple-choice setting, which are responsible for the answer decision and the format preference of the LMs, respectively. Then, we investigate the role of these two uncertainties on aligned LM's calibration through fine-tuning in simple synthetic alignment schemes and conclude that one reason for aligned LMs' overconfidence is the conflation of these two types of uncertainty. Furthermore, we examine the utility of common post-hoc calibration methods for aligned LMs and propose an easy-to-implement and sample-efficient method to calibrate aligned LMs. We hope our findings could provide insights into the design of more reliable alignment processes for LMs.
A fast normal splitting preconditioner for attractive coupled nonlinear Schrödinger equations with fractional Laplacian
Abstract
A linearly implicit conservative difference scheme is applied to discretize the attractive coupled nonlinear Schr\"odinger equations with fractional Laplacian. Complex symmetric linear systems can be obtained, and the system matrices are indefinite and Toeplitz-plus-diagonal. Neither efficient preconditioned iteration method nor fast direct method is available to deal with these systems. In this paper, we propose a novel matrix splitting iteration method based on a normal splitting of an equivalent real block form of the complex linear systems. This new iteration method converges unconditionally, and the quasi-optimal iteration parameter is deducted. The corresponding new preconditioner is obtained naturally, which can be constructed easily and implemented efficiently by fast Fourier transform. Theoretical analysis indicates that the eigenvalues of the preconditioned system matrix are tightly clustered. Numerical experiments show that the new preconditioner can significantly accelerate the convergence rate of the Krylov subspace iteration methods. Specifically, the convergence behavior of the related preconditioned GMRES iteration method is spacial mesh-size-independent, and almost fractional order insensitive. Moreover, the linearly implicit conservative difference scheme in conjunction with the preconditioned GMRES iteration method conserves the discrete mass and energy in terms of a given precision.
BanglaAbuseMeme: A Dataset for Bengali Abusive Meme Classification
Authors: Mithun Das, Animesh Mukherjee
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
The dramatic increase in the use of social media platforms for information sharing has also fueled a steep growth in online abuse. A simple yet effective way of abusing individuals or communities is by creating memes, which often integrate an image with a short piece of text layered on top of it. Such harmful elements are in rampant use and are a threat to online safety. Hence it is necessary to develop efficient models to detect and flag abusive memes. The problem becomes more challenging in a low-resource setting (e.g., Bengali memes, i.e., images with Bengali text embedded on it) because of the absence of benchmark datasets on which AI models could be trained. In this paper we bridge this gap by building a Bengali meme dataset. To setup an effective benchmark we implement several baseline models for classifying abusive memes using this dataset. We observe that multimodal models that use both textual and visual information outperform unimodal models. Our best-performing model achieves a macro F1 score of 70.51. Finally, we perform a qualitative error analysis of the misclassified memes of the best-performing text-based, image-based and multimodal models.
A Quasi-Wasserstein Loss for Learning Graph Neural Networks
Abstract
When learning graph neural networks (GNNs) in node-level prediction tasks, most existing loss functions are applied for each node independently, even if node embeddings and their labels are non-i.i.d. because of their graph structures. To eliminate such inconsistency, in this study we propose a novel Quasi-Wasserstein (QW) loss with the help of the optimal transport defined on graphs, leading to new learning and prediction paradigms of GNNs. In particular, we design a "Quasi-Wasserstein" distance between the observed multi-dimensional node labels and their estimations, optimizing the label transport defined on graph edges. The estimations are parameterized by a GNN in which the optimal label transport may determine the graph edge weights optionally. By reformulating the strict constraint of the label transport to a Bregman divergence-based regularizer, we obtain the proposed Quasi-Wasserstein loss associated with two efficient solvers learning the GNN together with optimal label transport. When predicting node labels, our model combines the output of the GNN with the residual component provided by the optimal label transport, leading to a new transductive prediction paradigm. Experiments show that the proposed QW loss applies to various GNNs and helps to improve their performance in node-level classification and regression tasks.
Annotated Job Ads with Named Entity Recognition
Authors: Felix Stollenwerk, Niklas Fastlund, Anna Nyqvist, Joey Öhman
Abstract
We have trained a named entity recognition (NER) model that screens Swedish job ads for different kinds of useful information (e.g. skills required from a job seeker). It was obtained by fine-tuning KB-BERT. The biggest challenge we faced was the creation of a labelled dataset, which required manual annotation. This paper gives an overview of the methods we employed to make the annotation process more efficient and to ensure high quality data. We also report on the performance of the resulting model.
Adversarial Training for Physics-Informed Neural Networks
Authors: Yao Li, Shengzhu Shi, Zhichang Guo, Boying Wu
Abstract
Physics-informed neural networks have shown great promise in solving partial differential equations. However, due to insufficient robustness, vanilla PINNs often face challenges when solving complex PDEs, especially those involving multi-scale behaviors or solutions with sharp or oscillatory characteristics. To address these issues, based on the projected gradient descent adversarial attack, we proposed an adversarial training strategy for PINNs termed by AT-PINNs. AT-PINNs enhance the robustness of PINNs by fine-tuning the model with adversarial samples, which can accurately identify model failure locations and drive the model to focus on those regions during training. AT-PINNs can also perform inference with temporal causality by selecting the initial collocation points around temporal initial values. We implement AT-PINNs to the elliptic equation with multi-scale coefficients, Poisson equation with multi-peak solutions, Burgers equation with sharp solutions and the Allen-Cahn equation. The results demonstrate that AT-PINNs can effectively locate and reduce failure regions. Moreover, AT-PINNs are suitable for solving complex PDEs, since locating failure regions through adversarial attacks is independent of the size of failure regions or the complexity of the distribution.
Abstract
The Internet is going through a massive infrastructural revolution with the advent of low-flying satellite networks, 5/6G, WiFi7, and hollow-core fiber deployments. While these networks could unleash enhanced connectivity and new capabilities, it is critical to understand the performance characteristics to efficiently drive applications over them. Low-Earth orbit (LEO) satellite mega-constellations like SpaceX Starlink aim to offer broad coverage and low latencies at the expense of high orbital dynamics leading to continuous latency changes and frequent satellite hand-offs. This paper aims to quantify Starlink's latency and its variations and components using a real testbed spanning multiple latitudes from the North to the South of Europe. We identify tail latencies as a problem. We develop predictors for latency and throughput and show their utility in improving application performance by up to 25%. We also explore how transport protocols can be optimized for LEO networks and show that this can improve throughput by up to 115% (with only a 5% increase in latency). Also, our measurement testbed with a footprint across multiple locations offers unique trigger-based scheduling capabilities that are necessary to quantify the impact of LEO dynamics.
Accelerate Presolve in Large-Scale Linear Programming via Reinforcement Learning
Abstract
Large-scale LP problems from industry usually contain much redundancy that severely hurts the efficiency and reliability of solving LPs, making presolve (i.e., the problem simplification module) one of the most critical components in modern LP solvers. However, how to design high-quality presolve routines -- that is, the program determining (P1) which presolvers to select, (P2) in what order to execute, and (P3) when to stop -- remains a highly challenging task due to the extensive requirements on expert knowledge and the large search space. Due to the sequential decision property of the task and the lack of expert demonstrations, we propose a simple and efficient reinforcement learning (RL) framework -- namely, reinforcement learning for presolve (RL4Presolve) -- to tackle (P1)-(P3) simultaneously. Specifically, we formulate the routine design task as a Markov decision process and propose an RL framework with adaptive action sequences to generate high-quality presolve routines efficiently. Note that adaptive action sequences help learn complex behaviors efficiently and adapt to various benchmarks. Experiments on two solvers (open-source and commercial) and eight benchmarks (real-world and synthetic) demonstrate that RL4Presolve significantly and consistently improves the efficiency of solving large-scale LPs, especially on benchmarks from industry. Furthermore, we optimize the hard-coded presolve routines in LP solvers by extracting rules from learned policies for simple and efficient deployment to Huawei's supply chain. The results show encouraging economic and academic potential for incorporating machine learning to modern solvers.
Sampling Algorithms for Butterfly Counting on Temporal Bipartite Graphs
Authors: Jiaxi Pu, Yanhao Wang, Yuchen Li, Xuan Zhou
Subjects: Social and Information Networks (cs.SI); Databases (cs.DB); Data Structures and Algorithms (cs.DS)
Abstract
Temporal bipartite graphs are widely used to denote time-evolving relationships between two disjoint sets of nodes, such as customer-product interactions in E-commerce and user-group memberships in social networks. Temporal butterflies, $(2,2)$-bicliques that occur within a short period and in a prescribed order, are essential in modeling the structural and sequential patterns of such graphs. Counting the number of temporal butterflies is thus a fundamental task in analyzing temporal bipartite graphs. However, existing algorithms for butterfly counting on static bipartite graphs and motif counting on temporal unipartite graphs are inefficient for this purpose. In this paper, we present a general framework with three sampling strategies for temporal butterfly counting. Since exact counting can be time-consuming on large graphs, our approach alternatively computes approximate estimates accurately and efficiently. We also provide analytical bounds on the number of samples each strategy requires to obtain estimates with small relative errors and high probability. We finally evaluate our framework on six real-world datasets and demonstrate its superior accuracy and efficiency compared to several baselines. Overall, our proposed framework and sampling strategies provide efficient and accurate approaches to approximating temporal butterfly counts on large-scale temporal bipartite graphs.
UNav-Sim: A Visually Realistic Underwater Robotics Simulator and Synthetic Data-generation Framework
Authors: Abdelhakim Amer, Olaya Álvarez-Tuñón, Halil Ibrahim Ugurlu, Jonas le Fevre Sejersen, Yury Brodskiy, Erdal Kayacan
Abstract
Underwater robotic surveys can be costly due to the complex working environment and the need for various sensor modalities. While underwater simulators are essential, many existing simulators lack sufficient rendering quality, restricting their ability to transfer algorithms from simulation to real-world applications. To address this limitation, we introduce UNav-Sim, which, to the best of our knowledge, is the first simulator to incorporate the efficient, high-detail rendering of Unreal Engine 5 (UE5). UNav-Sim is open-source and includes an autonomous vision-based navigation stack. By supporting standard robotics tools like ROS, UNav-Sim enables researchers to develop and test algorithms for underwater environments efficiently.
Simulating Users in Interactive Web Table Retrieval
Authors: Björn Engelmann, Timo Breuer, Philipp Schaer
Abstract
Considering the multimodal signals of search items is beneficial for retrieval effectiveness. Especially in web table retrieval (WTR) experiments, accounting for multimodal properties of tables boosts effectiveness. However, it still remains an open question how the single modalities affect user experience in particular. Previous work analyzed WTR performance in ad-hoc retrieval benchmarks, which neglects interactive search behavior and limits the conclusion about the implications for real-world user environments. To this end, this work presents an in-depth evaluation of simulated interactive WTR search sessions as a more cost-efficient and reproducible alternative to real user studies. As a first of its kind, we introduce interactive query reformulation strategies based on Doc2Query, incorporating cognitive states of simulated user knowledge. Our evaluations include two perspectives on user effectiveness by considering different cost paradigms, namely query-wise and time-oriented measures of effort. Our multi-perspective evaluation scheme reveals new insights about query strategies, the impact of modalities, and different user types in simulated WTR search sessions.
An EigenValue Stabilization Technique for Immersed Boundary Finite Element Methods in Explicit Dynamics
Authors: Sascha Eisenträger, Lars Radtke, Wadhah Garhuom, Stefan Löhnert, Alexander Düster, Daniel Juhre, Dominik Schillinger
Subjects: Numerical Analysis (math.NA); Dynamical Systems (math.DS)
Abstract
The application of immersed boundary methods in static analyses is often impeded by poorly cut elements (small cut elements problem), leading to ill-conditioned linear systems of equations and stability problems. While these concerns may not be paramount in explicit dynamics, a substantial reduction in the critical time step size based on the smallest volume fraction $\chi$ of a cut element is observed. This reduction can be so drastic that it renders explicit time integration schemes impractical. To tackle this challenge, we propose the use of a dedicated eigenvalue stabilization (EVS) technique. The EVS-technique serves a dual purpose. Beyond merely improving the condition number of system matrices, it plays a pivotal role in extending the critical time increment, effectively broadening the stability region in explicit dynamics. As a result, our approach enables robust and efficient analyses of high-frequency transient problems using immersed boundary methods. A key advantage of the stabilization method lies in the fact that only element-level operations are required. This is accomplished by computing all eigenvalues of the element matrices and subsequently introducing a stabilization term that mitigates the adverse effects of cutting. Notably, the stabilization of the mass matrix $\mathbf{M}\mathrm{c}$ of cut elements -- especially for high polynomial orders $p$ of the shape functions -- leads to a significant raise in the critical time step size $\Delta t\mathrm{cr}$. To demonstrate the efficacy of our technique, we present two specifically selected dynamic benchmark examples related to wave propagation analysis, where an explicit time integration scheme must be employed to leverage the increase in the critical time step size.
Abstract
In this work, we aim to establish a strong connection between two significant bodies of machine learning research: continual learning and sequence modeling. That is, we propose to formulate continual learning as a sequence modeling problem, allowing advanced sequence models to be utilized for continual learning. Under this formulation, the continual learning process becomes the forward pass of a sequence model. By adopting the meta-continual learning (MCL) framework, we can train the sequence model at the meta-level, on multiple continual learning episodes. As a specific example of our new formulation, we demonstrate the application of Transformers and their efficient variants as MCL methods. Our experiments on seven benchmarks, covering both classification and regression, show that sequence models can be an attractive solution for general MCL.
Acoustic shape optimization using energy stable curvilinear finite differences
Abstract
A gradient-based method for shape optimization problems constrained by the acoustic wave equation is presented. The method makes use of high-order accurate finite differences with summation-by-parts properties on multiblock curvilinear grids to discretize in space. Representing the design domain through a coordinate mapping from a reference domain, the design shape is obtained by inversion of the discretized coordinate map. The adjoint state framework is employed to efficiently compute the gradient of the loss functional. Using the summation-by-parts properties of the finite difference discretization, we prove stability and dual consistency for the semi-discrete forward and adjoint problems. Numerical experiments verify the accuracy of the finite difference scheme and demonstrate the capabilities of the shape optimization method on two model problems with real-world relevance.
Supporting UAVs with Edge Computing: A Review of Opportunities and Challenges
Authors: Malte Janßen, Tobias Pfandzelter, Minghe Wang, David Bermbach
Subjects: Distributed, Parallel, and Cluster Computing (cs.DC); Signal Processing (eess.SP)
Abstract
Over the last years, Unmanned Aerial Vehicles (UAVs) have seen significant advancements in sensor capabilities and computational abilities, allowing for efficient autonomous navigation and visual tracking applications. However, the demand for computationally complex tasks has increased faster than advances in battery technology. This opens up possibilities for improvements using edge computing. In edge computing, edge servers can achieve lower latency responses compared to traditional cloud servers through strategic geographic deployments. Furthermore, these servers can maintain superior computational performance compared to UAVs, as they are not limited by battery constraints. Combining these technologies by aiding UAVs with edge servers, research finds measurable improvements in task completion speed, energy efficiency, and reliability across multiple applications and industries. This systematic literature review aims to analyze the current state of research and collect, select, and extract the key areas where UAV activities can be supported and improved through edge computing.
Fast Multipole Attention: A Divide-and-Conquer Attention Mechanism for Long Sequences
Authors: Yanming Kang, Giang Tran, Hans De Sterck
Subjects: Computation and Language (cs.CL); Machine Learning (cs.LG)
Abstract
Transformer-based models have achieved state-of-the-art performance in many areas. However, the quadratic complexity of self-attention with respect to the input length hinders the applicability of Transformer-based models to long sequences. To address this, we present Fast Multipole Attention, a new attention mechanism that uses a divide-and-conquer strategy to reduce the time and memory complexity of attention for sequences of length $n$ from $\mathcal{O}(n^2)$ to $\mathcal{O}(n \log n)$ or $O(n)$, while retaining a global receptive field. The hierarchical approach groups queries, keys, and values into $\mathcal{O}( \log n)$ levels of resolution, where groups at greater distances are increasingly larger in size and the weights to compute group quantities are learned. As such, the interaction between tokens far from each other is considered in lower resolution in an efficient hierarchical manner. The overall complexity of Fast Multipole Attention is $\mathcal{O}(n)$ or $\mathcal{O}(n \log n)$, depending on whether the queries are down-sampled or not. This multi-level divide-and-conquer strategy is inspired by fast summation methods from $n$-body physics and the Fast Multipole Method. We perform evaluation on autoregressive and bidirectional language modeling tasks and compare our Fast Multipole Attention model with other efficient attention variants on medium-size datasets. We find empirically that the Fast Multipole Transformer performs much better than other efficient transformers in terms of memory size and accuracy. The Fast Multipole Attention mechanism has the potential to empower large language models with much greater sequence lengths, taking the full context into account in an efficient, naturally hierarchical manner during training and when generating long sequences.
Filling in the Gaps: Efficient Event Coreference Resolution using Graph Autoencoder Networks
Authors: Loic De Langhe, Orphée De Clercq, Veronique Hoste
Abstract
We introduce a novel and efficient method for Event Coreference Resolution (ECR) applied to a lower-resourced language domain. By framing ECR as a graph reconstruction task, we are able to combine deep semantic embeddings with structural coreference chain knowledge to create a parameter-efficient family of Graph Autoencoder models (GAE). Our method significantly outperforms classical mention-pair methods on a large Dutch event coreference corpus in terms of overall score, efficiency and training speed. Additionally, we show that our models are consistently able to classify more difficult coreference links and are far more robust in low-data settings when compared to transformer-based mention-pair coreference algorithms.
Subjects: Computational Engineering, Finance, and Science (cs.CE); Data Structures and Algorithms (cs.DS); Information Theory (cs.IT); Robotics (cs.RO); Systems and Control (eess.SY)
Abstract
In this paper we present a framework of key algorithms and data-structures for efficiently generating timetables for any number of AGVs from any given positioning on any given graph to accomplish any given demands as long as a few easily satisfiable assumptions are met. Our proposed algorithms provide guaranteed solutions in predictable polynomial running-times, which is fundamental to any real-time application. We also develop an improved geographic reservation algorithm that provides a substantial run-time improvement of the previously best-known algorithm from $O(nm)$ to $O(n)$.
Exact and efficient solutions of the LMC Multitask Gaussian Process model
Abstract
The Linear Model of Co-regionalization (LMC) is a very general model of multitask gaussian process for regression or classification. While its expressivity and conceptual simplicity are appealing, naive implementations have cubic complexity in the number of datapoints and number of tasks, making approximations mandatory for most applications. However, recent work has shown that under some conditions the latent processes of the model can be decoupled, leading to a complexity that is only linear in the number of said processes. We here extend these results, showing from the most general assumptions that the only condition necessary to an efficient exact computation of the LMC is a mild hypothesis on the noise model. We introduce a full parametrization of the resulting \emph{projected LMC} model, and an expression of the marginal likelihood enabling efficient optimization. We perform a parametric study on synthetic data to show the excellent performance of our approach, compared to an unrestricted exact LMC and approximations of the latter. Overall, the projected LMC appears as a credible and simpler alternative to state-of-the art models, which greatly facilitates some computations such as leave-one-out cross-validation and fantasization.
A General Theoretical Paradigm to Understand Learning from Human Preferences
Authors: Mohammad Gheshlaghi Azar, Mark Rowland, Bilal Piot, Daniel Guo, Daniele Calandriello, Michal Valko, Rémi Munos
Abstract
The prevalent deployment of learning from human preferences through reinforcement learning (RLHF) relies on two important approximations: the first assumes that pairwise preferences can be substituted with pointwise rewards. The second assumes that a reward model trained on these pointwise rewards can generalize from collected data to out-of-distribution data sampled by the policy. Recently, Direct Preference Optimisation (DPO) has been proposed as an approach that bypasses the second approximation and learn directly a policy from collected data without the reward modelling stage. However, this method still heavily relies on the first approximation. In this paper we try to gain a deeper theoretical understanding of these practical algorithms. In particular we derive a new general objective called $\Psi$PO for learning from human preferences that is expressed in terms of pairwise preferences and therefore bypasses both approximations. This new general objective allows us to perform an in-depth analysis of the behavior of RLHF and DPO (as special cases of $\Psi$PO) and to identify their potential pitfalls. We then consider another special case for $\Psi$PO by setting $\Psi$ simply to Identity, for which we can derive an efficient optimisation procedure, prove performance guarantees and demonstrate its empirical superiority to DPO on some illustrative examples.
Ordered Reliability Direct Error Pattern Testing Decoding Algorithm
Authors: Reza Hadavian, Xiaoting Huang, Dmitri Truhachev, Kamal El-Sankary, Hamid Ebrahimzad, Hossein Najafi
Subjects: Information Theory (cs.IT); Signal Processing (eess.SP)
Abstract
We introduce a novel universal soft-decision decoding algorithm for binary block codes called ordered reliability direct error pattern testing (ORDEPT). Our results, obtained for a variety of popular short high-rate codes, demonstrate that ORDEPT outperforms state-of-the-art decoding algorithms of comparable complexity such as ordered reliability bits guessing random additive noise decoding (ORBGRAND) in terms of the decoding error probability and latency. The improvements carry on to the iterative decoding of product codes and convolutional product-like codes, where we present a new adaptive decoding algorithm and demonstrate the ability of ORDEPT to efficiently find multiple candidate codewords to produce soft output.
Machine Learning-based Nutrient Application's Timeline Recommendation for Smart Agriculture: A Large-Scale Data Mining Approach
Abstract
This study addresses the vital role of data analytics in monitoring fertiliser applications in crop cultivation. Inaccurate fertiliser application decisions can lead to costly consequences, hinder food production, and cause environmental harm. We propose a solution to predict nutrient application by determining required fertiliser quantities for an entire season. The proposed solution recommends adjusting fertiliser amounts based on weather conditions and soil characteristics to promote cost-effective and environmentally friendly agriculture. The collected dataset is high-dimensional and heterogeneous. Our research examines large-scale heterogeneous datasets in the context of the decision-making process, encompassing data collection and analysis. We also study the impact of fertiliser applications combined with weather data on crop yield, using the winter wheat crop as a case study. By understanding local contextual and geographic factors, we aspire to stabilise or even reduce the demand for agricultural nutrients while enhancing crop development. The proposed approach is proven to be efficient and scalable, as it is validated using a real-world and large dataset.
Rational function approximation with normalized positive denominators
Abstract
Rational function approximations provide a simple but flexible alternative to polynomial approximation, allowing one to capture complex non-linearities without oscillatory artifacts. However, there have been few attempts to use rational functions on noisy data due to the likelihood of creating spurious singularities. To avoid the creation of singularities, we use Bernstein polynomials and appropriate conditions on their coefficients to force the denominator to be strictly positive. While this reduces the range of rational polynomials that can be expressed, it keeps all the benefits of rational functions while maintaining the robustness of polynomial approximation in noisy data scenarios. Our numerical experiments on noisy data show that existing rational approximation methods continually produce spurious poles inside the approximation domain. This contrasts our method, which cannot create poles in the approximation domain and provides better fits than a polynomial approximation and even penalized splines on functions with multiple variables. Moreover, guaranteeing pole-free in an interval is critical for estimating non-constant coefficients when numerically solving differential equations using spectral methods. This provides a compact representation of the original differential equation, allowing numeric solvers to achieve high accuracy quickly, as seen in our experiments.
Robust Class-Conditional Distribution Alignment for Partial Domain Adaptation
Authors: Sandipan Choudhuri, Arunabha Sen
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Unwanted samples from private source categories in the learning objective of a partial domain adaptation setup can lead to negative transfer and reduce classification performance. Existing methods, such as re-weighting or aggregating target predictions, are vulnerable to this issue, especially during initial training stages, and do not adequately address overlapping categorical distributions. We propose a solution to overcome these limitations by exploring beyond the first-order moments for robust alignment of categorical distributions. We employ objectives that optimize the intra and inter-class distributions in a domain-invariant fashion and design a robust pseudo-labeling for efficient target supervision. Our approach incorporates a complement entropy objective module to reduce classification uncertainty and flatten incorrect category predictions. The experimental findings and ablation analysis of the proposed modules demonstrate the superior performance of our proposed model compared to benchmarks.
Black-Box Training Data Identification in GANs via Detector Networks
Abstract
Since their inception Generative Adversarial Networks (GANs) have been popular generative models across images, audio, video, and tabular data. In this paper we study whether given access to a trained GAN, as well as fresh samples from the underlying distribution, if it is possible for an attacker to efficiently identify if a given point is a member of the GAN's training data. This is of interest for both reasons related to copyright, where a user may want to determine if their copyrighted data has been used to train a GAN, and in the study of data privacy, where the ability to detect training set membership is known as a membership inference attack. Unlike the majority of prior work this paper investigates the privacy implications of using GANs in black-box settings, where the attack only has access to samples from the generator, rather than access to the discriminator as well. We introduce a suite of membership inference attacks against GANs in the black-box setting and evaluate our attacks on image GANs trained on the CIFAR10 dataset and tabular GANs trained on genomic data. Our most successful attack, called The Detector, involve training a second network to score samples based on their likelihood of being generated by the GAN, as opposed to a fresh sample from the distribution. We prove under a simple model of the generator that the detector is an approximately optimal membership inference attack. Across a wide range of tabular and image datasets, attacks, and GAN architectures, we find that adversaries can orchestrate non-trivial privacy attacks when provided with access to samples from the generator. At the same time, the attack success achievable against GANs still appears to be lower compared to other generative and discriminative models; this leaves the intriguing open question of whether GANs are in fact more private, or if it is a matter of developing stronger attacks.
Non-Intrusive Adaptation: Input-Centric Parameter-efficient Fine-Tuning for Versatile Multimodal Modeling
Authors: Yaqing Wang, Jialin Wu, Tanmaya Dabral, Jiageng Zhang, Geoff Brown, Chun-Ta Lu, Frederick Liu, Yi Liang, Bo Pang, Michael Bendersky, Radu Soricut
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Multimedia (cs.MM)
Abstract
Large language models (LLMs) and vision language models (VLMs) demonstrate excellent performance on a wide range of tasks by scaling up parameter counts from O(10^9) to O(10^{12}) levels and further beyond. These large scales make it impossible to adapt and deploy fully specialized models given a task of interest. Parameter-efficient fine-tuning (PEFT) emerges as a promising direction to tackle the adaptation and serving challenges for such large models. We categorize PEFT techniques into two types: intrusive and non-intrusive. Intrusive PEFT techniques directly change a model's internal architecture. Though more flexible, they introduce significant complexities for training and serving. Non-intrusive PEFT techniques leave the internal architecture unchanged and only adapt model-external parameters, such as embeddings for input. In this work, we describe AdaLink as a non-intrusive PEFT technique that achieves competitive performance compared to SoTA intrusive PEFT (LoRA) and full model fine-tuning (FT) on various tasks. We evaluate using both text-only and multimodal tasks, with experiments that account for both parameter-count scaling and training regime (with and without instruction tuning).
CAPGrasp: An $\mathbb{R}^3\times \text{SO(2)-equivariant}$ Continuous Approach-Constrained Generative Grasp Sampler
Abstract
We propose CAPGrasp, an $\mathbb{R}^3\times \text{SO(2)-equivariant}$ 6-DoF continuous approach-constrained generative grasp sampler. It includes a novel learning strategy for training CAPGrasp that eliminates the need to curate massive conditionally labeled datasets and a constrained grasp refinement technique that improves grasp poses while respecting the grasp approach directional constraints. The experimental results demonstrate that CAPGrasp is more than three times as sample efficient as unconstrained grasp samplers while achieving up to 38% grasp success rate improvement. CAPGrasp also achieves 4-10% higher grasp success rates than constrained but noncontinuous grasp samplers. Overall, CAPGrasp is a sample-efficient solution when grasps must originate from specific directions, such as grasping in confined spaces.
Distributed Indexing Schemes for k-Dominant Skyline Analytics on Uncertain Edge-IoT Data
Authors: Chuan-Chi Lai, Hsuan-Yu Lin, Chuan-Ming Liu
Subjects: Distributed, Parallel, and Cluster Computing (cs.DC); Databases (cs.DB)
Abstract
Skyline queries typically search a Pareto-optimal set from a given data set to solve the corresponding multiobjective optimization problem. As the number of criteria increases, the skyline presumes excessive data items, which yield a meaningless result. To address this curse of dimensionality, we proposed a k-dominant skyline in which the number of skyline members was reduced by relaxing the restriction on the number of dimensions, considering the uncertainty of data. Specifically, each data item was associated with a probability of appearance, which represented the probability of becoming a member of the k-dominant skyline. As data items appear continuously in data streams, the corresponding k-dominant skyline may vary with time. Therefore, an effective and rapid mechanism of updating the k-dominant skyline becomes crucial. Herein, we proposed two time-efficient schemes, Middle Indexing (MI) and All Indexing (AI), for k-dominant skyline in distributed edge-computing environments, where irrelevant data items can be effectively excluded from the compute to reduce the processing duration. Furthermore, the proposed schemes were validated with extensive experimental simulations. The experimental results demonstrated that the proposed MI and AI schemes reduced the computation time by approximately 13% and 56%, respectively, compared with the existing method.
SHARCS: Efficient Transformers through Routing with Dynamic Width Sub-networks
Abstract
We introduce SHARCS for adaptive inference that takes into account the hardness of input samples. SHARCS can train a router on any transformer network, enabling the model to direct different samples to sub-networks with varying widths. Our experiments demonstrate that: (1) SHARCS outperforms or complements existing per-sample adaptive inference methods across various classification tasks in terms of accuracy vs. FLOPs; (2) SHARCS generalizes across different architectures and can be even applied to compressed and efficient transformer encoders to further improve their efficiency; (3) SHARCS can provide a 2 times inference speed up at an insignificant drop in accuracy.
Keyword: faster
Contracting Tsetlin Machine with Absorbing Automata
Authors: Bimal Bhattarai, Ole-Christoffer Granmo, Lei Jiao, Per-Arne Andersen, Svein Anders Tunheim, Rishad Shafik, Alex Yakovlev
Subjects: Formal Languages and Automata Theory (cs.FL); Artificial Intelligence (cs.AI)
Abstract
In this paper, we introduce a sparse Tsetlin Machine (TM) with absorbing Tsetlin Automata (TA) states. In brief, the TA of each clause literal has both an absorbing Exclude- and an absorbing Include state, making the learning scheme absorbing instead of ergodic. When a TA reaches an absorbing state, it will never leave that state again. If the absorbing state is an Exclude state, both the automaton and the literal can be removed from further consideration. The literal will as a result never participates in that clause. If the absorbing state is an Include state, on the other hand, the literal is stored as a permanent part of the clause while the TA is discarded. A novel sparse data structure supports these updates by means of three action lists: Absorbed Include, Include, and Exclude. By updating these lists, the TM gets smaller and smaller as the literals and their TA withdraw. In this manner, the computation accelerates during learning, leading to faster learning and less energy consumption.
Learning a Hierarchical Planner from Humans in Multiple Generations
Authors: Leonardo Hernandez Cano, Yewen Pu, Robert D. Hawkins, Josh Tenenbaum, Armando Solar-Lezama
Abstract
A typical way in which a machine acquires knowledge from humans is by programming. Compared to learning from demonstrations or experiences, programmatic learning allows the machine to acquire a novel skill as soon as the program is written, and, by building a library of programs, a machine can quickly learn how to perform complex tasks. However, as programs often take their execution contexts for granted, they are brittle when the contexts change, making it difficult to adapt complex programs to new contexts. We present natural programming, a library learning system that combines programmatic learning with a hierarchical planner. Natural programming maintains a library of decompositions, consisting of a goal, a linguistic description of how this goal decompose into sub-goals, and a concrete instance of its decomposition into sub-goals. A user teaches the system via curriculum building, by identifying a challenging yet not impossible goal along with linguistic hints on how this goal may be decomposed into sub-goals. The system solves for the goal via hierarchical planning, using the linguistic hints to guide its probability distribution in proposing the right plans. The system learns from this interaction by adding newly found decompositions in the successful search into its library. Simulated studies and a human experiment (n=360) on a controlled environment demonstrate that natural programming can robustly compose programs learned from different users and contexts, adapting faster and solving more complex tasks when compared to programmatic baselines.
Estimating Material Properties of Interacting Objects Using Sum-GP-UCB
Abstract
Robots need to estimate the material and dynamic properties of objects from observations in order to simulate them accurately. We present a Bayesian optimization approach to identifying the material property parameters of objects based on a set of observations. Our focus is on estimating these properties based on observations of scenes with different sets of interacting objects. We propose an approach that exploits the structure of the reward function by modeling the reward for each observation separately and using only the parameters of the objects in that scene as inputs. The resulting lower-dimensional models generalize better over the parameter space, which in turn results in a faster optimization. To speed up the optimization process further, and reduce the number of simulation runs needed to find good parameter values, we also propose partial evaluations of the reward function, wherein the selected parameters are only evaluated on a subset of real world evaluations. The approach was successfully evaluated on a set of scenes with a wide range of object interactions, and we showed that our method can effectively perform incremental learning without resetting the rewards of the gathered observations.
Learning to Generate Parameters of ConvNets for Unseen Image Data
Authors: Shiye Wang, Kaituo Feng, Changsheng Li, Ye Yuan, Guoren Wang
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Typical Convolutional Neural Networks (ConvNets) depend heavily on large amounts of image data and resort to an iterative optimization algorithm (e.g., SGD or Adam) to learn network parameters, which makes training very time- and resource-intensive. In this paper, we propose a new training paradigm and formulate the parameter learning of ConvNets into a prediction task: given a ConvNet architecture, we observe there exists correlations between image datasets and their corresponding optimal network parameters, and explore if we can learn a hyper-mapping between them to capture the relations, such that we can directly predict the parameters of the network for an image dataset never seen during the training phase. To do this, we put forward a new hypernetwork based model, called PudNet, which intends to learn a mapping between datasets and their corresponding network parameters, and then predicts parameters for unseen data with only a single forward propagation. Moreover, our model benefits from a series of adaptive hyper recurrent units sharing weights to capture the dependencies of parameters among different network layers. Extensive experiments demonstrate that our proposed method achieves good efficacy for unseen image datasets on two kinds of settings: Intra-dataset prediction and Inter-dataset prediction. Our PudNet can also well scale up to large-scale datasets, e.g., ImageNet-1K. It takes 8967 GPU seconds to train ResNet-18 on the ImageNet-1K using GC from scratch and obtain a top-5 accuracy of 44.65 %. However, our PudNet costs only 3.89 GPU seconds to predict the network parameters of ResNet-18 achieving comparable performance (44.92 %), more than 2,300 times faster than the traditional training paradigm.
Accelerated Policy Gradient: On the Nesterov Momentum for Reinforcement Learning
Abstract
Policy gradient methods have recently been shown to enjoy global convergence at a $\Theta(1/t)$ rate in the non-regularized tabular softmax setting. Accordingly, one important research question is whether this convergence rate can be further improved, with only first-order updates. In this paper, we answer the above question from the perspective of momentum by adapting the celebrated Nesterov's accelerated gradient (NAG) method to reinforcement learning (RL), termed \textit{Accelerated Policy Gradient} (APG). To demonstrate the potential of APG in achieving faster global convergence, we formally show that with the true gradient, APG with softmax policy parametrization converges to an optimal policy at a $\tilde{O}(1/t^2)$ rate. To the best of our knowledge, this is the first characterization of the global convergence rate of NAG in the context of RL. Notably, our analysis relies on one interesting finding: Regardless of the initialization, APG could end up reaching a locally nearly-concave regime, where APG could benefit significantly from the momentum, within finite iterations. By means of numerical validation, we confirm that APG exhibits $\tilde{O}(1/t^2)$ rate as well as show that APG could significantly improve the convergence behavior over the standard policy gradient.
Supporting UAVs with Edge Computing: A Review of Opportunities and Challenges
Authors: Malte Janßen, Tobias Pfandzelter, Minghe Wang, David Bermbach
Subjects: Distributed, Parallel, and Cluster Computing (cs.DC); Signal Processing (eess.SP)
Abstract
Over the last years, Unmanned Aerial Vehicles (UAVs) have seen significant advancements in sensor capabilities and computational abilities, allowing for efficient autonomous navigation and visual tracking applications. However, the demand for computationally complex tasks has increased faster than advances in battery technology. This opens up possibilities for improvements using edge computing. In edge computing, edge servers can achieve lower latency responses compared to traditional cloud servers through strategic geographic deployments. Furthermore, these servers can maintain superior computational performance compared to UAVs, as they are not limited by battery constraints. Combining these technologies by aiding UAVs with edge servers, research finds measurable improvements in task completion speed, energy efficiency, and reliability across multiple applications and industries. This systematic literature review aims to analyze the current state of research and collect, select, and extract the key areas where UAV activities can be supported and improved through edge computing.
Keyword: mobile
WaveFlex: A Smart Surface for Private CBRS Wireless Cellular Networks
Authors: Fan Yi, Kun Woo Cho, Yaxiong Xie, Kyle Jamieson
Subjects: Networking and Internet Architecture (cs.NI); Signal Processing (eess.SP)
Abstract
We present the design and implementation of WaveFlex, the first smart surface that enhances Private LTE/5G networks operating under the shared-license framework in the Citizens Broadband Radio Service frequency band. WaveFlex works in the presence of frequency diversity: multiple nearby base stations operating on different frequencies, as dictated by a Spectrum Access System coordinator. It also handles time dynamism: due to the dynamic sharing rules of the band, base stations occasionally switch channels, especially when priority users enter the network. Finally, WaveFlex operates independently of the network itself, not requiring access to nor modification of the base station or mobile users, yet it remain compliant with and effective on prevailing cellular protocols. We have designed and fabricated WaveFlex on a custom multi-layer PCB, software defined radio-based network monitor, and supporting control software and hardware. Our experimental evaluation benchmarks an operational Private LTE network running at full line rate. Results demonstrate an 8.50 dB average SNR gain, and an average throughput gain of 4.36 Mbps for a single small cell, and 3.19 Mbps for four small cells, in a realistic indoor office scenario.
Towards Inferring Users' Impressions of Robot Performance in Navigation Scenarios
Authors: Qiping Zhang, Nathan Tsoi, Booyeon Choi, Jie Tan, Hao-Tien Lewis Chiang, Marynel Vázquez
Abstract
Human impressions of robot performance are often measured through surveys. As a more scalable and cost-effective alternative, we study the possibility of predicting people's impressions of robot behavior using non-verbal behavioral cues and machine learning techniques. To this end, we first contribute the SEAN TOGETHER Dataset consisting of observations of an interaction between a person and a mobile robot in a Virtual Reality simulation, together with impressions of robot performance provided by users on a 5-point scale. Second, we contribute analyses of how well humans and supervised learning techniques can predict perceived robot performance based on different combinations of observation types (e.g., facial, spatial, and map features). Our results show that facial expressions alone provide useful information about human impressions of robot performance; but in the navigation scenarios we tested, spatial features are the most critical piece of information for this inference task. Also, when evaluating results as binary classification (rather than multiclass classification), the F1-Score of human predictions and machine learning models more than doubles, showing that both are better at telling the directionality of robot performance than predicting exact performance ratings. Based on our findings, we provide guidelines for implementing these predictions models in real-world navigation scenarios.
Forward Kinematics of Object Transport by a Multi-Robot System with Deformable Sheet
Abstract
We present object handling and transport by a multi-robot team with a deformable sheet as a carrier. Due to the deformability of the sheet and the high dimension of the whole system, it is challenging to clearly describe all the possible positions of the object on the sheet for a given formation of the multi-robot system. A complete forward kinematics (FK) method is proposed in this paper for object handling by an $N$-mobile robot team with a deformable sheet. Based on the virtual variable cables model, a constrained quadratic problem (CQP) is formulated by combining the form closure and minimum potential energy conditions of the system. Analytical solutions to the CQP are presented and then further verified with the force closure condition. With the proposed FK method, all possible solutions are obtained with the given initial sheet shape and the robot team formation. We demonstrate the effectiveness, completeness, and efficiency of the FK method with simulation and experimental results.
Min-max Decoding Error Probability Optimization in RIS-Aided Hybrid TDMA-NOMA Networks
Authors: Tra Huong Thi Le, Yan Kyaw Tun
Subjects: Networking and Internet Architecture (cs.NI); Information Theory (cs.IT)
Abstract
One of the primary objectives for future wireless communication networks is to facilitate the provision of ultra-reliable and low-latency communication services while simultaneously ensuring the capability for vast connection. In order to achieve this objective, we examine a hybrid multi-access scheme inside the finite blocklength (FBL) regime. This system combines the benefits of non-orthogonal multiple access (NOMA) and time-division multiple access (TDMA) schemes with the aim of fulfilling the objectives of future wireless communication networks. In addition, a reconfigurable intelligent surface (RIS) is utilized to facilitate the establishment of the uplink transmission between the base station and mobile devices in situations when impediments impede their direct communication linkages. This paper aims to minimize the worst-case decoding-error probability for all mobile users by jointly optimizing power allocation, receiving beamforming, blocklength, RIS reflection, and user pairing. To deal with the coupled variables in the formulated mixed-integer non-convex optimization problem, we decompose it into three sub-problems, namely, 1) decoding order determination problem, 2) joint power allocation, receiving beamforming, RIS reflection, and blocklength optimization problem, and 3) optimal user pairing problem. Then, we provide the sequential convex approximation (SCA) and semidefinite relaxation (SDR)-based algorithms as potential solutions for iteratively addressing the deconstructed first two sub-problems at a fixed random user pairing. In addition, the Hungarian matching approach is employed to address the challenge of optimizing user pairing. In conclusion, we undertake a comprehensive simulation, which reveals the advantageous qualities of the proposed algorithm and its superior performance compared to existing benchmark methods.
Do We Run Large-scale Multi-Robot Systems on the Edge? More Evidence for Two-Phase Performance in System Size Scaling
Authors: Jonas Kuckling, Robin Luckey, Viktor Avrutin, Andrew Vardy, Andreagiovanni Reina, Heiko Hamann
Subjects: Robotics (cs.RO); Multiagent Systems (cs.MA)
Abstract
With increasing numbers of mobile robots arriving in real-world applications, more robots coexist in the same space, interact, and possibly collaborate. Methods to provide such systems with system size scalability are known, for example, from swarm robotics. Example strategies are self-organizing behavior, a strict decentralized approach, and limiting the robot-robot communication. Despite applying such strategies, any multi-robot system breaks above a certain critical system size (i.e., number of robots) as too many robots share a resource (e.g., space, communication channel). We provide additional evidence based on simulations, that at these critical system sizes, the system performance separates into two phases: nearly optimal and minimal performance. We speculate that in real-world applications that are configured for optimal system size, the supposedly high-performing system may actually live on borrowed time as it is on a transient to breakdown. We provide two modeling options (based on queueing theory and a population model) that may help to support this reasoning.
Take the aTrain. Introducing an Interface for the Accessible Transcription of Interviews
Authors: Armin Haberl, Jürgen Fleiß, Dominik Kowald, Stefan Thalmann
Abstract
aTrain is an open-source and offline tool for transcribing audio data in multiple languages with CPU and NVIDIA GPU support. It is specifically designed for researchers using qualitative data generated from various forms of speech interactions with research participants. aTrain requires no programming skills, runs on most computers, does not require an internet connection, and was verified not to upload data to any server. aTrain combines OpenAI's Whisper model with speaker recognition to provide output that integrates with the popular qualitative data analysis software tools MAXQDA and ATLAS.ti. It has an easy-to-use graphical interface and is provided as a Windows-App through the Microsoft Store allowing for simple installation by researchers. The source code is freely available on GitHub. Having developed aTrain with a focus on speed on local computers, we show that the transcription time on current mobile CPUs is around 2 to 3 times the duration of the audio file using the highest-accuracy transcription models. If an entry-level graphics card is available, the transcription speed increases to 20% of the audio duration.
Keyword: pruning
In defense of parameter sharing for model-compression
Abstract
When considering a model architecture, there are several ways to reduce its memory footprint. Historically, popular approaches included selecting smaller architectures and creating sparse networks through pruning. More recently, randomized parameter-sharing (RPS) methods have gained traction for model compression at start of training. In this paper, we comprehensively assess the trade-off between memory and accuracy across RPS, pruning techniques, and building smaller models. Our findings demonstrate that RPS, which is both data and model-agnostic, consistently outperforms/matches smaller models and all moderately informed pruning strategies, such as MAG, SNIP, SYNFLOW, and GRASP, across the entire compression range. This advantage becomes particularly pronounced in higher compression scenarios. Notably, even when compared to highly informed pruning techniques like Lottery Ticket Rewinding (LTR), RPS exhibits superior performance in high compression settings. This points out inherent capacity advantage that RPS enjoys over sparse models. Theoretically, we establish RPS as a superior technique in terms of memory-efficient representation when compared to pruning for linear models. This paper argues in favor of paradigm shift towards RPS based models. During our rigorous evaluation of RPS, we identified issues in the state-of-the-art RPS technique ROAST, specifically regarding stability (ROAST's sensitivity to initialization hyperparameters, often leading to divergence) and Pareto-continuity (ROAST's inability to recover the accuracy of the original model at zero compression). We provably address both of these issues. We refer to the modified RPS, which incorporates our improvements, as STABLE-RPS.
Abstract
Price movements in financial markets are well known to be very noisy. As a result, even if there are, on occasion, exploitable patterns that could be picked up by machine-learning algorithms, these are obscured by feature and label noise rendering the predictions less useful, and risky in practice. Traditional rule-learning techniques developed for noisy data, such as CN2, would seek only high precision rules and refrain from making predictions where their antecedents did not apply. We apply a similar approach, where a model abstains from making a prediction on data points that it is uncertain on. During training, a cascade of such models are learned in sequence, similar to rule lists, with each model being trained only on data on which the previous model(s) were uncertain. Similar pruning of data takes place at test-time, with (higher accuracy) predictions being made albeit only on a fraction (support) of test-time data. In a financial prediction setting, such an approach allows decisions to be taken only when the ensemble model is confident, thereby reducing risk. We present results using traditional MLPs as well as differentiable decision trees, on synthetic data as well as real financial market data, to predict fixed-term returns using commonly used features. We submit that our approach is likely to result in better overall returns at a lower level of risk. In this context we introduce an utility metric to measure the average gain per trade, as well as the return adjusted for downside risk, both of which are improved significantly by our approach.
Abstract
Accurately forecasting the motion of traffic actors is crucial for the deployment of autonomous vehicles at a large scale. Current trajectory forecasting approaches primarily concentrate on optimizing a loss function with a specific metric, which can result in predictions that do not adhere to physical laws or violate external constraints. Our objective is to incorporate explicit knowledge priors that allow a network to forecast future trajectories in compliance with both the kinematic constraints of a vehicle and the geometry of the driving environment. To achieve this, we introduce a non-parametric pruning layer and attention layers to integrate the defined knowledge priors. Our proposed method is designed to ensure reachability guarantees for traffic actors in both complex and dynamic situations. By conditioning the network to follow physical laws, we can obtain accurate and safe predictions, essential for maintaining autonomous vehicles' safety and efficiency in real-world settings.In summary, this paper presents concepts that prevent off-road predictions for safe and reliable motion forecasting by incorporating knowledge priors into the training process.
Keyword: diffusion
GenEval: An Object-Focused Framework for Evaluating Text-to-Image Alignment
Authors: Dhruba Ghosh, Hanna Hajishirzi, Ludwig Schmidt
Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Abstract
Recent breakthroughs in diffusion models, multimodal pretraining, and efficient finetuning have led to an explosion of text-to-image generative models. Given human evaluation is expensive and difficult to scale, automated methods are critical for evaluating the increasingly large number of new models. However, most current automated evaluation metrics like FID or CLIPScore only offer a holistic measure of image quality or image-text alignment, and are unsuited for fine-grained or instance-level analysis. In this paper, we introduce GenEval, an object-focused framework to evaluate compositional image properties such as object co-occurrence, position, count, and color. We show that current object detection models can be leveraged to evaluate text-to-image models on a variety of generation tasks with strong human agreement, and that other discriminative vision models can be linked to this pipeline to further verify properties like object color. We then evaluate several open-source text-to-image models and analyze their relative generative capabilities on our benchmark. We find that recent models demonstrate significant improvement on these tasks, though they are still lacking in complex capabilities such as spatial relations and attribute binding. Finally, we demonstrate how GenEval might be used to help discover existing failure modes, in order to inform development of the next generation of text-to-image models. Our code to run the GenEval framework is publicly available at https://github.com/djghosh13/geneval.
Reflection-Equivariant Diffusion for 3D Structure Determination from Isotopologue Rotational Spectra in Natural Abundance
Abstract
Structure determination is necessary to identify unknown organic molecules, such as those in natural products, forensic samples, the interstellar medium, and laboratory syntheses. Rotational spectroscopy enables structure determination by providing accurate 3D information about small organic molecules via their moments of inertia. Using these moments, Kraitchman analysis determines isotopic substitution coordinates, which are the unsigned $|x|,|y|,|z|$ coordinates of all atoms with natural isotopic abundance, including carbon, nitrogen, and oxygen. While unsigned substitution coordinates can verify guesses of structures, the missing $+/-$ signs make it challenging to determine the actual structure from the substitution coordinates alone. To tackle this inverse problem, we develop KREED (Kraitchman REflection-Equivariant Diffusion), a generative diffusion model that infers a molecule's complete 3D structure from its molecular formula, moments of inertia, and unsigned substitution coordinates of heavy atoms. KREED's top-1 predictions identify the correct 3D structure with >98% accuracy on the QM9 and GEOM datasets when provided with substitution coordinates of all heavy atoms with natural isotopic abundance. When substitution coordinates are restricted to only a subset of carbons, accuracy is retained at 91% on QM9 and 32% on GEOM. On a test set of experimentally measured substitution coordinates gathered from the literature, KREED predicts the correct all-atom 3D structure in 25 of 33 cases, demonstrating experimental applicability for context-free 3D structure determination with rotational spectroscopy.
On the Evaluation of Generative Models in Distributed Learning Tasks
Abstract
The evaluation of deep generative models including generative adversarial networks (GANs) and diffusion models has been extensively studied in the literature. While the existing evaluation methods mainly target a centralized learning problem with training data stored by a single client, many applications of generative models concern distributed learning settings, e.g. the federated learning scenario, where training data are collected by and distributed among several clients. In this paper, we study the evaluation of generative models in distributed learning tasks with heterogeneous data distributions. First, we focus on the Fr\'echet inception distance (FID) and consider the following FID-based aggregate scores over the clients: 1) FID-avg as the mean of clients' individual FID scores, 2) FID-all as the FID distance of the trained model to the collective dataset containing all clients' data. We prove that the model rankings according to the FID-all and FID-avg scores could be inconsistent, which can lead to different optimal generative models according to the two aggregate scores. Next, we consider the kernel inception distance (KID) and similarly define the KID-avg and KID-all aggregations. Unlike the FID case, we prove that KID-all and KID-avg result in the same rankings of generative models. We perform several numerical experiments on standard image datasets and training schemes to support our theoretical findings on the evaluation of generative models in distributed learning problems.
Language Agents for Detecting Implicit Stereotypes in Text-to-image Models at Scale
Authors: Qichao Wang, Tian Bian, Yian Yin, Tingyang Xu, Hong Cheng, Helen M. Meng, Zibin Zheng, Liang Chen, Bingzhe Wu
Subjects: Computers and Society (cs.CY); Computation and Language (cs.CL)
Abstract
The recent surge in the research of diffusion models has accelerated the adoption of text-to-image models in various Artificial Intelligence Generated Content (AIGC) commercial products. While these exceptional AIGC products are gaining increasing recognition and sparking enthusiasm among consumers, the questions regarding whether, when, and how these models might unintentionally reinforce existing societal stereotypes remain largely unaddressed. Motivated by recent advancements in language agents, here we introduce a novel agent architecture tailored for stereotype detection in text-to-image models. This versatile agent architecture is capable of accommodating free-form detection tasks and can autonomously invoke various tools to facilitate the entire process, from generating corresponding instructions and images, to detecting stereotypes. We build the stereotype-relevant benchmark based on multiple open-text datasets, and apply this architecture to commercial products and popular open source text-to-image models. We find that these models often display serious stereotypes when it comes to certain prompts about personal characteristics, social cultural context and crime-related aspects. In summary, these empirical findings underscore the pervasive existence of stereotypes across social dimensions, including gender, race, and religion, which not only validate the effectiveness of our proposed approach, but also emphasize the critical necessity of addressing potential ethical risks in the burgeoning realm of AIGC. As AIGC continues its rapid expansion trajectory, with new models and plugins emerging daily in staggering numbers, the challenge lies in the timely detection and mitigation of potential biases within these models.
Progressive3D: Progressively Local Editing for Text-to-3D Content Creation with Complex Semantic Prompts
Authors: Xinhua Cheng, Tianyu Yang, Jianan Wang, Yu Li, Lei Zhang, Jian Zhang, Li Yuan
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Recent text-to-3D generation methods achieve impressive 3D content creation capacity thanks to the advances in image diffusion models and optimizing strategies. However, current methods struggle to generate correct 3D content for a complex prompt in semantics, i.e., a prompt describing multiple interacted objects binding with different attributes. In this work, we propose a general framework named Progressive3D, which decomposes the entire generation into a series of locally progressive editing steps to create precise 3D content for complex prompts, and we constrain the content change to only occur in regions determined by user-defined region prompts in each editing step. Furthermore, we propose an overlapped semantic component suppression technique to encourage the optimization process to focus more on the semantic differences between prompts. Extensive experiments demonstrate that the proposed Progressive3D framework generates precise 3D content for prompts with complex semantics and is general for various text-to-3D methods driven by different 3D representations.
De novo protein design using geometric vector field networks
Abstract
Innovations like protein diffusion have enabled significant progress in de novo protein design, which is a vital topic in life science. These methods typically depend on protein structure encoders to model residue backbone frames, where atoms do not exist. Most prior encoders rely on atom-wise features, such as angles and distances between atoms, which are not available in this context. Thus far, only several simple encoders, such as IPA, have been proposed for this scenario, exposing the frame modeling as a bottleneck. In this work, we proffer the Vector Field Network (VFN), which enables network layers to perform learnable vector computations between coordinates of frame-anchored virtual atoms, thus achieving a higher capability for modeling frames. The vector computation operates in a manner similar to a linear layer, with each input channel receiving 3D virtual atom coordinates instead of scalar values. The multiple feature vectors output by the vector computation are then used to update the residue representations and virtual atom coordinates via attention aggregation. Remarkably, VFN also excels in modeling both frames and atoms, as the real atoms can be treated as the virtual atoms for modeling, positioning VFN as a potential universal encoder. In protein diffusion (frame modeling), VFN exhibits an impressive performance advantage over IPA, excelling in terms of both designability (67.04% vs. 53.58%) and diversity (66.54% vs. 51.98%). In inverse folding (frame and atom modeling), VFN outperforms the previous SoTA model, PiFold (54.7% vs. 51.66%), on sequence recovery rate. We also propose a method of equipping VFN with the ESM model, which significantly surpasses the previous ESM-based SoTA (62.67% vs. 55.65%), LM-Design, by a substantial margin.
To Generate or Not? Safety-Driven Unlearned Diffusion Models Are Still Easy To Generate Unsafe Images ... For Now
Authors: Yimeng Zhang, Jinghan Jia, Xin Chen, Aochuan Chen, Yihua Zhang, Jiancheng Liu, Ke Ding, Sijia Liu
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
The recent advances in diffusion models (DMs) have revolutionized the generation of complex and diverse images. However, these models also introduce potential safety hazards, such as the production of harmful content and infringement of data copyrights. Although there have been efforts to create safety-driven unlearning methods to counteract these challenges, doubts remain about their capabilities. To bridge this uncertainty, we propose an evaluation framework built upon adversarial attacks (also referred to as adversarial prompts), in order to discern the trustworthiness of these safety-driven unlearned DMs. Specifically, our research explores the (worst-case) robustness of unlearned DMs in eradicating unwanted concepts, styles, and objects, assessed by the generation of adversarial prompts. We develop a novel adversarial learning approach called UnlearnDiff that leverages the inherent classification capabilities of DMs to streamline the generation of adversarial prompts, making it as simple for DMs as it is for image classification attacks. This technique streamlines the creation of adversarial prompts, making the process as intuitive for generative modeling as it is for image classification assaults. Through comprehensive benchmarking, we assess the unlearning robustness of five prevalent unlearned DMs across multiple tasks. Our results underscore the effectiveness and efficiency of UnlearnDiff when compared to state-of-the-art adversarial prompting methods. Codes are available at https://github.com/OPTML-Group/Diffusion-MU-Attack. WARNING: This paper contains model outputs that may be offensive in nature.
IRAD: Implicit Representation-driven Image Resampling against Adversarial Attacks
Abstract
We introduce a novel approach to counter adversarial attacks, namely, image resampling. Image resampling transforms a discrete image into a new one, simulating the process of scene recapturing or rerendering as specified by a geometrical transformation. The underlying rationale behind our idea is that image resampling can alleviate the influence of adversarial perturbations while preserving essential semantic information, thereby conferring an inherent advantage in defending against adversarial attacks. To validate this concept, we present a comprehensive study on leveraging image resampling to defend against adversarial attacks. We have developed basic resampling methods that employ interpolation strategies and coordinate shifting magnitudes. Our analysis reveals that these basic methods can partially mitigate adversarial attacks. However, they come with apparent limitations: the accuracy of clean images noticeably decreases, while the improvement in accuracy on adversarial examples is not substantial. We propose implicit representation-driven image resampling (IRAD) to overcome these limitations. First, we construct an implicit continuous representation that enables us to represent any input image within a continuous coordinate space. Second, we introduce SampleNet, which automatically generates pixel-wise shifts for resampling in response to different inputs. Furthermore, we can extend our approach to the state-of-the-art diffusion-based method, accelerating it with fewer time steps while preserving its defense capability. Extensive experiments demonstrate that our method significantly enhances the adversarial robustness of diverse deep models against various attacks while maintaining high accuracy on clean images.
InfoDiffusion: Information Entropy Aware Diffusion Process for Non-Autoregressive Text Generation
Authors: Renzhi Wang, Jing Li, Piji Li
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
Abstract
Diffusion models have garnered considerable interest in the field of text generation. Several studies have explored text diffusion models with different structures and applied them to various tasks, including named entity recognition and summarization. However, there exists a notable disparity between the "easy-first" text generation process of current diffusion models and the "keyword-first" natural text generation process of humans, which has received limited attention. To bridge this gap, we propose InfoDiffusion, a non-autoregressive text diffusion model. Our approach introduces a "keyinfo-first" generation strategy and incorporates a noise schedule based on the amount of text information. In addition, InfoDiffusion combines self-conditioning with a newly proposed partially noising model structure. Experimental results show that InfoDiffusion outperforms the baseline model in terms of generation quality and diversity, as well as exhibiting higher sampling efficiency.
Image Super-resolution Via Latent Diffusion: A Sampling-space Mixture Of Experts And Frequency-augmented Decoder Approach
Authors: Feng Luo, Jinxi Xiang, Jun Zhang, Xiao Han, Wei Yang
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
The recent use of diffusion prior, enhanced by pre-trained text-image models, has markedly elevated the performance of image super-resolution (SR). To alleviate the huge computational cost required by pixel-based diffusion SR, latent-based methods utilize a feature encoder to transform the image and then implement the SR image generation in a compact latent space. Nevertheless, there are two major issues that limit the performance of latent-based diffusion. First, the compression of latent space usually causes reconstruction distortion. Second, huge computational cost constrains the parameter scale of the diffusion model. To counteract these issues, we first propose a frequency compensation module that enhances the frequency components from latent space to pixel space. The reconstruction distortion (especially for high-frequency information) can be significantly decreased. Then, we propose to use Sample-Space Mixture of Experts (SS-MoE) to achieve more powerful latent-based SR, which steadily improves the capacity of the model without a significant increase in inference costs. These carefully crafted designs contribute to performance improvements in largely explored 4x blind super-resolution benchmarks and extend to large magnification factors, i.e., 8x image SR benchmarks. The code is available at https://github.com/amandaluof/moe_sr.
Quality Diversity through Human Feedback
Authors: Li Ding, Jenny Zhang, Jeff Clune, Lee Spector, Joel Lehman
Subjects: Artificial Intelligence (cs.AI); Neural and Evolutionary Computing (cs.NE)
Abstract
Reinforcement learning from human feedback (RLHF) has exhibited the potential to enhance the performance of foundation models for qualitative tasks. Despite its promise, its efficacy is often restricted when conceptualized merely as a mechanism to maximize learned reward models of averaged human preferences, especially in areas such as image generation which demand diverse model responses. Meanwhile, quality diversity (QD) algorithms, dedicated to seeking diverse, high-quality solutions, are often constrained by the dependency on manually defined diversity metrics. Interestingly, such limitations of RLHF and QD can be overcome by blending insights from both. This paper introduces Quality Diversity through Human Feedback (QDHF), which employs human feedback for inferring diversity metrics, expanding the applicability of QD algorithms. Empirical results reveal that QDHF outperforms existing QD methods regarding automatic diversity discovery, and matches the search capabilities of QD with human-constructed metrics. Notably, when deployed for a latent space illumination task, QDHF markedly enhances the diversity of images generated by a Diffusion model. The study concludes with an in-depth analysis of QDHF's sample efficiency and the quality of its derived diversity metrics, emphasizing its promise for enhancing exploration and diversity in optimization for complex, open-ended tasks.
Keyword: adaptive
Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection
Abstract
Despite their remarkable capabilities, large language models (LLMs) often produce responses containing factual inaccuracies due to their sole reliance on the parametric knowledge they encapsulate. Retrieval-Augmented Generation (RAG), an ad hoc approach that augments LMs with retrieval of relevant knowledge, decreases such issues. However, indiscriminately retrieving and incorporating a fixed number of retrieved passages, regardless of whether retrieval is necessary, or passages are relevant, diminishes LM versatility or can lead to unhelpful response generation. We introduce a new framework called Self-Reflective Retrieval-Augmented Generation (Self-RAG) that enhances an LM's quality and factuality through retrieval and self-reflection. Our framework trains a single arbitrary LM that adaptively retrieves passages on-demand, and generates and reflects on retrieved passages and its own generations using special tokens, called reflection tokens. Generating reflection tokens makes the LM controllable during the inference phase, enabling it to tailor its behavior to diverse task requirements. Experiments show that Self-RAG (7B and 13B parameters) significantly outperforms state-of-the-art LLMs and retrieval-augmented models on a diverse set of tasks. Specifically, Self-RAG outperforms ChatGPT and retrieval-augmented Llama2-chat on Open-domain QA, reasoning and fact verification tasks, and it shows significant gains in improving factuality and citation accuracy for long-form generations relative to these models.
Value-Biased Maximum Likelihood Estimation for Model-based Reinforcement Learning in Discounted Linear MDPs
Authors: Yu-Heng Hung, Ping-Chun Hsieh, Akshay Mete, P. R. Kumar
Abstract
We consider the infinite-horizon linear Markov Decision Processes (MDPs), where the transition probabilities of the dynamic model can be linearly parameterized with the help of a predefined low-dimensional feature mapping. While the existing regression-based approaches have been theoretically shown to achieve nearly-optimal regret, they are computationally rather inefficient due to the need for a large number of optimization runs in each time step, especially when the state and action spaces are large. To address this issue, we propose to solve linear MDPs through the lens of Value-Biased Maximum Likelihood Estimation (VBMLE), which is a classic model-based exploration principle in the adaptive control literature for resolving the well-known closed-loop identification problem of Maximum Likelihood Estimation. We formally show that (i) VBMLE enjoys $\widetilde{O}(d\sqrt{T})$ regret, where $T$ is the time horizon and $d$ is the dimension of the model parameter, and (ii) VBMLE is computationally more efficient as it only requires solving one optimization problem in each time step. In our regret analysis, we offer a generic convergence result of MLE in linear MDPs through a novel supermartingale construct and uncover an interesting connection between linear MDPs and online learning, which could be of independent interest. Finally, the simulation results show that VBMLE significantly outperforms the benchmark method in terms of both empirical regret and computation time.
Differentially Private Data Generation with Missing Data
Authors: Shubhankar Mohapatra, Jianqiao Zong, Florian Kerschbaum, Xi He
Subjects: Databases (cs.DB); Cryptography and Security (cs.CR)
Abstract
Despite several works that succeed in generating synthetic data with differential privacy (DP) guarantees, they are inadequate for generating high-quality synthetic data when the input data has missing values. In this work, we formalize the problems of DP synthetic data with missing values and propose three effective adaptive strategies that significantly improve the utility of the synthetic data on four real-world datasets with different types and levels of missing data and privacy requirements. We also identify the relationship between privacy impact for the complete ground truth data and incomplete data for these DP synthetic data generation algorithms. We model the missing mechanisms as a sampling process to obtain tighter upper bounds for the privacy guarantees to the ground truth data. Overall, this study contributes to a better understanding of the challenges and opportunities for using private synthetic data generation algorithms in the presence of missing data.
WaveAttack: Asymmetric Frequency Obfuscation-based Backdoor Attacks Against Deep Neural Networks
Abstract
Due to the popularity of Artificial Intelligence (AI) technology, numerous backdoor attacks are designed by adversaries to mislead deep neural network predictions by manipulating training samples and training processes. Although backdoor attacks are effective in various real scenarios, they still suffer from the problems of both low fidelity of poisoned samples and non-negligible transfer in latent space, which make them easily detectable by existing backdoor detection algorithms. To overcome the weakness, this paper proposes a novel frequency-based backdoor attack method named WaveAttack, which obtains image high-frequency features through Discrete Wavelet Transform (DWT) to generate backdoor triggers. Furthermore, we introduce an asymmetric frequency obfuscation method, which can add an adaptive residual in the training and inference stage to improve the impact of triggers and further enhance the effectiveness of WaveAttack. Comprehensive experimental results show that WaveAttack not only achieves higher stealthiness and effectiveness, but also outperforms state-of-the-art (SOTA) backdoor attack methods in the fidelity of images by up to 28.27\% improvement in PSNR, 1.61\% improvement in SSIM, and 70.59\% reduction in IS. Our code is available at https://anonymous.4open.science/r/AnonymousRep-701D.
MISAR: A Multimodal Instructional System with Augmented Reality
Authors: Jing Bi, Nguyen Manh Nguyen, Ali Vosoughi, Chenliang Xu
Subjects: Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
Abstract
Augmented reality (AR) requires the seamless integration of visual, auditory, and linguistic channels for optimized human-computer interaction. While auditory and visual inputs facilitate real-time and contextual user guidance, the potential of large language models (LLMs) in this landscape remains largely untapped. Our study introduces an innovative method harnessing LLMs to assimilate information from visual, auditory, and contextual modalities. Focusing on the unique challenge of task performance quantification in AR, we utilize egocentric video, speech, and context analysis. The integration of LLMs facilitates enhanced state estimation, marking a step towards more adaptive AR systems. Code, dataset, and demo will be available at https://github.com/nguyennm1024/misar.
Action-Quantized Offline Reinforcement Learning for Robotic Skill Learning
Abstract
The offline reinforcement learning (RL) paradigm provides a general recipe to convert static behavior datasets into policies that can perform better than the policy that collected the data. While policy constraints, conservatism, and other methods for mitigating distributional shifts have made offline reinforcement learning more effective, the continuous action setting often necessitates various approximations for applying these techniques. Many of these challenges are greatly alleviated in discrete action settings, where offline RL constraints and regularizers can often be computed more precisely or even exactly. In this paper, we propose an adaptive scheme for action quantization. We use a VQ-VAE to learn state-conditioned action quantization, avoiding the exponential blowup that comes with na\"ive discretization of the action space. We show that several state-of-the-art offline RL methods such as IQL, CQL, and BRAC improve in performance on benchmarks when combined with our proposed discretization scheme. We further validate our approach on a set of challenging long-horizon complex robotic manipulation tasks in the Robomimic environment, where our discretized offline RL algorithms are able to improve upon their continuous counterparts by 2-3x. Our project page is at https://saqrl.github.io/
AdaVis: Adaptive and Explainable Visualization Recommendation for Tabular Data
Authors: Songheng Zhang, Haotian Li, Huamin Qu, Yong Wang
Abstract
Automated visualization recommendation facilitates the rapid creation of effective visualizations, which is especially beneficial for users with limited time and limited knowledge of data visualization. There is an increasing trend in leveraging machine learning (ML) techniques to achieve an end-to-end visualization recommendation. However, existing ML-based approaches implicitly assume that there is only one appropriate visualization for a specific dataset, which is often not true for real applications. Also, they often work like a black box, and are difficult for users to understand the reasons for recommending specific visualizations. To fill the research gap, we propose AdaVis, an adaptive and explainable approach to recommend one or multiple appropriate visualizations for a tabular dataset. It leverages a box embedding-based knowledge graph to well model the possible one-to-many mapping relations among different entities (i.e., data features, dataset columns, datasets, and visualization choices). The embeddings of the entities and relations can be learned from dataset-visualization pairs. Also, AdaVis incorporates the attention mechanism into the inference framework. Attention can indicate the relative importance of data features for a dataset and provide fine-grained explainability. Our extensive evaluations through quantitative metric evaluations, case studies, and user interviews demonstrate the effectiveness of AdaVis.
Hierarchical accompanying and inhibiting patterns on the spatial arrangement of taxis' local hotspots
Authors: Xiao-Jian Chen, Changjiang Xiao, Zhou Huanga, Keli Wang, Weiyu Zhang, Yu Liu
Abstract
Due to the large volume of recording, the complete spontaneity, and the flexible pick-up and drop-off locations, taxi data portrays a realistic and detailed picture of urban space use to a certain extent. The spatial arrangement of pick-up and drop-off hotspots reflects the organizational space, which has received attention in urban structure studies. Previous studies mainly explore the hotspots at a large scale by visual analysis or some simple indexes, where the hotspots usually cover the entire central business district, train stations, or dense residential areas, reaching a radius of hundreds or even thousands of meters. However, the spatial arrangement patterns of small-scale hotspots, reflecting the specific popular pick-up and drop-off locations, have not received much attention. Using two taxi trajectory datasets in Wuhan and Beijing, China, this study quantitatively explores the spatial arrangement of fine-grained pick-up and drop-off local hotspots with different levels of popularity, where the sizes are adaptively set as 90m90m in Wuhan and 105m105m in Beijing according to the local hotspot identification method. Results show that popular hotspots tend to be surrounded by less popular hotspots, but the existence of less popular hotspots is inhibited in regions with a large number of popular hotspots. We use the terms hierarchical accompany and inhibiting patterns for these two spatial configurations. Finally, to uncover the underlying mechanism, a KNN-based model is proposed to reproduce the spatial distribution of other less popular hotspots according to the most popular ones. These findings help decision-makers construct reasonable urban minimum units for precise traffic and disease control, as well as plan a more humane spatial arrangement of points of interest.
Accelerate Presolve in Large-Scale Linear Programming via Reinforcement Learning
Abstract
Large-scale LP problems from industry usually contain much redundancy that severely hurts the efficiency and reliability of solving LPs, making presolve (i.e., the problem simplification module) one of the most critical components in modern LP solvers. However, how to design high-quality presolve routines -- that is, the program determining (P1) which presolvers to select, (P2) in what order to execute, and (P3) when to stop -- remains a highly challenging task due to the extensive requirements on expert knowledge and the large search space. Due to the sequential decision property of the task and the lack of expert demonstrations, we propose a simple and efficient reinforcement learning (RL) framework -- namely, reinforcement learning for presolve (RL4Presolve) -- to tackle (P1)-(P3) simultaneously. Specifically, we formulate the routine design task as a Markov decision process and propose an RL framework with adaptive action sequences to generate high-quality presolve routines efficiently. Note that adaptive action sequences help learn complex behaviors efficiently and adapt to various benchmarks. Experiments on two solvers (open-source and commercial) and eight benchmarks (real-world and synthetic) demonstrate that RL4Presolve significantly and consistently improves the efficiency of solving large-scale LPs, especially on benchmarks from industry. Furthermore, we optimize the hard-coded presolve routines in LP solvers by extracting rules from learned policies for simple and efficient deployment to Huawei's supply chain. The results show encouraging economic and academic potential for incorporating machine learning to modern solvers.
Learning to Generate Parameters of ConvNets for Unseen Image Data
Authors: Shiye Wang, Kaituo Feng, Changsheng Li, Ye Yuan, Guoren Wang
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Typical Convolutional Neural Networks (ConvNets) depend heavily on large amounts of image data and resort to an iterative optimization algorithm (e.g., SGD or Adam) to learn network parameters, which makes training very time- and resource-intensive. In this paper, we propose a new training paradigm and formulate the parameter learning of ConvNets into a prediction task: given a ConvNet architecture, we observe there exists correlations between image datasets and their corresponding optimal network parameters, and explore if we can learn a hyper-mapping between them to capture the relations, such that we can directly predict the parameters of the network for an image dataset never seen during the training phase. To do this, we put forward a new hypernetwork based model, called PudNet, which intends to learn a mapping between datasets and their corresponding network parameters, and then predicts parameters for unseen data with only a single forward propagation. Moreover, our model benefits from a series of adaptive hyper recurrent units sharing weights to capture the dependencies of parameters among different network layers. Extensive experiments demonstrate that our proposed method achieves good efficacy for unseen image datasets on two kinds of settings: Intra-dataset prediction and Inter-dataset prediction. Our PudNet can also well scale up to large-scale datasets, e.g., ImageNet-1K. It takes 8967 GPU seconds to train ResNet-18 on the ImageNet-1K using GC from scratch and obtain a top-5 accuracy of 44.65 %. However, our PudNet costs only 3.89 GPU seconds to predict the network parameters of ResNet-18 achieving comparable performance (44.92 %), more than 2,300 times faster than the traditional training paradigm.
Stochastic Optimization for Non-convex Problem with Inexact Hessian Matrix, Gradient, and Function
Authors: Liu Liu, Xuanqing Liu, Cho-Jui Hsieh, Dacheng Tao
Abstract
Trust-region (TR) and adaptive regularization using cubics (ARC) have proven to have some very appealing theoretical properties for non-convex optimization by concurrently computing function value, gradient, and Hessian matrix to obtain the next search direction and the adjusted parameters. Although stochastic approximations help largely reduce the computational cost, it is challenging to theoretically guarantee the convergence rate. In this paper, we explore a family of stochastic TR and ARC methods that can simultaneously provide inexact computations of the Hessian matrix, gradient, and function values. Our algorithms require much fewer propagations overhead per iteration than TR and ARC. We prove that the iteration complexity to achieve $\epsilon$-approximate second-order optimality is of the same order as the exact computations demonstrated in previous studies. Additionally, the mild conditions on inexactness can be met by leveraging a random sampling technology in the finite-sum minimization problem. Numerical experiments with a non-convex problem support these findings and demonstrate that, with the same or a similar number of iterations, our algorithms require less computational overhead per iteration than current second-order methods.
Abstract
The broad topic of this thesis is the design and analysis of Bitcoin custody systems. Both the technology and threat landscape are evolving constantly. Therefore, custody systems, defence strategies, and risk models should be adaptive too. We introduce Bitcoin custody by describing the different types, design principles, phases and functions of custody systems. We review the technology stack of these systems and focus on the fundamentals; key-management and privacy. We present a perspective we call the systems view. It is an attempt to capture the full complexity of a custody system, including technology, people, and processes. We review existing custody systems and standards. We explore Bitcoin covenants. This is a mechanism to enforce constraints on transaction sequences. Although previous work has proposed how to construct and apply Bitcoin covenants, these require modifying the consensus rules of Bitcoin, a notoriously difficult task. We introduce the first detailed exposition and security analysis of a deleted-key covenant protocol, which is compatible with current consensus rules. We demonstrate a range of security models for deleted-key covenants which seem practical, in particular, when applied in autonomous (user-controlled) custody systems. We conclude with a comparative analysis with previous proposals. Covenants are often proclaimed to be an important primitive for custody systems, but no complete design has been proposed to validate that claim. To address this, we propose an autonomous custody system called Ajolote which uses deleted-key covenants to enforce a vault sequence. We evaluate Ajolote with; a model of its state dynamics, a privacy analysis, and a risk model. We propose a threat model for custody systems which captures a realistic attacker for a system with offline devices and user-verification. We perform ceremony analysis to construct the risk model.
Improving Generalization of Alignment with Human Preferences through Group Invariant Learning
Abstract
The success of AI assistants based on language models (LLMs) hinges crucially on Reinforcement Learning from Human Feedback (RLHF), which enables the generation of responses more aligned with human preferences. As universal AI assistants, there's a growing expectation for them to perform consistently across various domains. However, previous work shows that Reinforcement Learning (RL) often exploits shortcuts to attain high rewards and overlooks challenging samples. This focus on quick reward gains undermines both the stability in training and the model's ability to generalize to new, unseen data. In this work, we propose a novel approach that can learn a consistent policy via RL across various data groups or domains. Given the challenges associated with acquiring group annotations, our method automatically classifies data into different groups, deliberately maximizing performance variance. Then, we optimize the policy to perform well on challenging groups. Lastly, leveraging the established groups, our approach adaptively adjusts the exploration space, allocating more learning capacity to more challenging data and preventing the model from over-optimizing on simpler data. Experimental results indicate that our approach significantly enhances training stability and model generalization.
SegmATRon: Embodied Adaptive Semantic Segmentation for Indoor Environment
Authors: Tatiana Zemskova, Margarita Kichik, Dmitry Yudin, Aleksei Staroverov, Aleksandr Panov
Abstract
This paper presents an adaptive transformer model named SegmATRon for embodied image semantic segmentation. Its distinctive feature is the adaptation of model weights during inference on several images using a hybrid multicomponent loss function. We studied this model on datasets collected in the photorealistic Habitat and the synthetic AI2-THOR Simulators. We showed that obtaining additional images using the agent's actions in an indoor environment can improve the quality of semantic segmentation. The code of the proposed approach and datasets are publicly available at https://github.com/wingrune/SegmATRon.
Ordered Reliability Direct Error Pattern Testing Decoding Algorithm
Authors: Reza Hadavian, Xiaoting Huang, Dmitri Truhachev, Kamal El-Sankary, Hamid Ebrahimzad, Hossein Najafi
Subjects: Information Theory (cs.IT); Signal Processing (eess.SP)
Abstract
We introduce a novel universal soft-decision decoding algorithm for binary block codes called ordered reliability direct error pattern testing (ORDEPT). Our results, obtained for a variety of popular short high-rate codes, demonstrate that ORDEPT outperforms state-of-the-art decoding algorithms of comparable complexity such as ordered reliability bits guessing random additive noise decoding (ORBGRAND) in terms of the decoding error probability and latency. The improvements carry on to the iterative decoding of product codes and convolutional product-like codes, where we present a new adaptive decoding algorithm and demonstrate the ability of ORDEPT to efficiently find multiple candidate codewords to produce soft output.
SHARCS: Efficient Transformers through Routing with Dynamic Width Sub-networks
Abstract
We introduce SHARCS for adaptive inference that takes into account the hardness of input samples. SHARCS can train a router on any transformer network, enabling the model to direct different samples to sub-networks with varying widths. Our experiments demonstrate that: (1) SHARCS outperforms or complements existing per-sample adaptive inference methods across various classification tasks in terms of accuracy vs. FLOPs; (2) SHARCS generalizes across different architectures and can be even applied to compressed and efficient transformer encoders to further improve their efficiency; (3) SHARCS can provide a 2 times inference speed up at an insignificant drop in accuracy.
Keyword: quantization
A Comprehensive Survey on Vector Database: Storage and Retrieval Technique, Challenge
Abstract
A vector database is used to store high-dimensional data that cannot be characterized by traditional DBMS. Although there are not many articles describing existing or introducing new vector database architectures, the approximate nearest neighbor search problem behind vector databases has been studied for a long time, and considerable related algorithmic articles can be found in the literature. This article attempts to comprehensively review relevant algorithms to provide a general understanding of this booming research area. The basis of our framework categorises these studies by the approach of solving ANNS problem, respectively hash-based, tree-based, graph-based and quantization-based approaches. Then we present an overview of existing challenges for vector databases. Lastly, we sketch how vector databases can be combined with large language models and provide new possibilities.
Action-Quantized Offline Reinforcement Learning for Robotic Skill Learning
Abstract
The offline reinforcement learning (RL) paradigm provides a general recipe to convert static behavior datasets into policies that can perform better than the policy that collected the data. While policy constraints, conservatism, and other methods for mitigating distributional shifts have made offline reinforcement learning more effective, the continuous action setting often necessitates various approximations for applying these techniques. Many of these challenges are greatly alleviated in discrete action settings, where offline RL constraints and regularizers can often be computed more precisely or even exactly. In this paper, we propose an adaptive scheme for action quantization. We use a VQ-VAE to learn state-conditioned action quantization, avoiding the exponential blowup that comes with na\"ive discretization of the action space. We show that several state-of-the-art offline RL methods such as IQL, CQL, and BRAC improve in performance on benchmarks when combined with our proposed discretization scheme. We further validate our approach on a set of challenging long-horizon complex robotic manipulation tasks in the Robomimic environment, where our discretized offline RL algorithms are able to improve upon their continuous counterparts by 2-3x. Our project page is at https://saqrl.github.io/
VQ-NeRF: Neural Reflectance Decomposition and Editing with Vector Quantization
Abstract
We propose VQ-NeRF, a two-branch neural network model that incorporates Vector Quantization (VQ) to decompose and edit reflectance fields in 3D scenes. Conventional neural reflectance fields use only continuous representations to model 3D scenes, despite the fact that objects are typically composed of discrete materials in reality. This lack of discretization can result in noisy material decomposition and complicated material editing. To address these limitations, our model consists of a continuous branch and a discrete branch. The continuous branch follows the conventional pipeline to predict decomposed materials, while the discrete branch uses the VQ mechanism to quantize continuous materials into individual ones. By discretizing the materials, our model can reduce noise in the decomposition process and generate a segmentation map of discrete materials. Specific materials can be easily selected for further editing by clicking on the corresponding area of the segmentation outcomes. Additionally, we propose a dropout-based VQ codeword ranking strategy to predict the number of materials in a scene, which reduces redundancy in the material segmentation process. To improve usability, we also develop an interactive interface to further assist material editing. We evaluate our model on both computer-generated and real-world scenes, demonstrating its superior performance. To the best of our knowledge, our model is the first to enable discrete material editing in 3D scenes.
Keyword: efficient
ASP: Automatic Selection of Proxy dataset for efficient AutoML
On the Temperature of Bayesian Graph Neural Networks for Conformal Prediction
GenEval: An Object-Focused Framework for Evaluating Text-to-Image Alignment
Value-Biased Maximum Likelihood Estimation for Model-based Reinforcement Learning in Discounted Linear MDPs
Efficient Online Learning with Offline Datasets for Infinite Horizon MDPs: A Bayesian Approach
HMN: Generalization of Heterogeneous and Multi-layered Network
Towards Optimal Regret in Adversarial Linear MDPs with Bandit Feedback
Personalized Soups: Personalized Large Language Model Alignment via Post-hoc Parameter Merging
FreSh: A Lock-Free Data Series Index
In defense of parameter sharing for model-compression
VKIE: The Application of Key Information Extraction on Video Text
US Microelectronics Packaging Ecosystem: Challenges and Opportunities
High Efficiency Polymer based Direct Multi-jet Impingement Cooling Solution for High Power Devices
Prototype-based HyperAdapter for Sample-Efficient Multi-task Tuning
MixEdit: Revisiting Data Augmentation and Beyond for Grammatical Error Correction
PREM: A Simple Yet Effective Approach for Node-Level Graph Anomaly Detection
Improved Sample Complexity Analysis of Natural Policy Gradient Algorithm with General Parameterization for Infinite Horizon Discounted Reward Markov Decision Processes
Descriptive Knowledge Graph in Biomedical Domain
Quantum Acceleration of Infinite Horizon Average-Reward Reinforcement Learning
Superiority of Softmax: Unveiling the Performance Edge Over Linear Attention
Adaptation with Self-Evaluation to Improve Selective Prediction in LLMs
Runner re-identification from single-view video in the open-world setting
DPF-Nutrition: Food Nutrition Estimation via Depth Prediction and Fusion
Quantify Health-Related Atomic Knowledge in Chinese Medical Large Language Models: A Computational Analysis
VST++: Efficient and Stronger Visual Saliency Transformer
EchoScan: Scanning Complex Indoor Geometries via Acoustic Echoes
Investigating Uncertainty Calibration of Aligned Language Models under the Multiple-Choice Setting
A fast normal splitting preconditioner for attractive coupled nonlinear Schrödinger equations with fractional Laplacian
BanglaAbuseMeme: A Dataset for Bengali Abusive Meme Classification
A Quasi-Wasserstein Loss for Learning Graph Neural Networks
Annotated Job Ads with Named Entity Recognition
Adversarial Training for Physics-Informed Neural Networks
T3P: Demystifying Low-Earth Orbit Satellite Broadband
Accelerate Presolve in Large-Scale Linear Programming via Reinforcement Learning
Sampling Algorithms for Butterfly Counting on Temporal Bipartite Graphs
UNav-Sim: A Visually Realistic Underwater Robotics Simulator and Synthetic Data-generation Framework
Simulating Users in Interactive Web Table Retrieval
An EigenValue Stabilization Technique for Immersed Boundary Finite Element Methods in Explicit Dynamics
Recasting Continual Learning as Sequence Modeling
Acoustic shape optimization using energy stable curvilinear finite differences
Supporting UAVs with Edge Computing: A Review of Opportunities and Challenges
Fast Multipole Attention: A Divide-and-Conquer Attention Mechanism for Long Sequences
Filling in the Gaps: Efficient Event Coreference Resolution using Graph Autoencoder Networks
Guaranteed, Predictable, Polynomial AGV Time-Pathing
Exact and efficient solutions of the LMC Multitask Gaussian Process model
A General Theoretical Paradigm to Understand Learning from Human Preferences
Ordered Reliability Direct Error Pattern Testing Decoding Algorithm
Machine Learning-based Nutrient Application's Timeline Recommendation for Smart Agriculture: A Large-Scale Data Mining Approach
Rational function approximation with normalized positive denominators
Robust Class-Conditional Distribution Alignment for Partial Domain Adaptation
Black-Box Training Data Identification in GANs via Detector Networks
Non-Intrusive Adaptation: Input-Centric Parameter-efficient Fine-Tuning for Versatile Multimodal Modeling
CAPGrasp: An $\mathbb{R}^3\times \text{SO(2)-equivariant}$ Continuous Approach-Constrained Generative Grasp Sampler
Distributed Indexing Schemes for k-Dominant Skyline Analytics on Uncertain Edge-IoT Data
SHARCS: Efficient Transformers through Routing with Dynamic Width Sub-networks
Keyword: faster
Contracting Tsetlin Machine with Absorbing Automata
Learning a Hierarchical Planner from Humans in Multiple Generations
Estimating Material Properties of Interacting Objects Using Sum-GP-UCB
Learning to Generate Parameters of ConvNets for Unseen Image Data
Accelerated Policy Gradient: On the Nesterov Momentum for Reinforcement Learning
Supporting UAVs with Edge Computing: A Review of Opportunities and Challenges
Keyword: mobile
WaveFlex: A Smart Surface for Private CBRS Wireless Cellular Networks
Towards Inferring Users' Impressions of Robot Performance in Navigation Scenarios
Forward Kinematics of Object Transport by a Multi-Robot System with Deformable Sheet
Min-max Decoding Error Probability Optimization in RIS-Aided Hybrid TDMA-NOMA Networks
Do We Run Large-scale Multi-Robot Systems on the Edge? More Evidence for Two-Phase Performance in System Size Scaling
Take the aTrain. Introducing an Interface for the Accessible Transcription of Interviews
Keyword: pruning
In defense of parameter sharing for model-compression
Conservative Predictions on Noisy Financial Data
KI-PMF: Knowledge Integrated Plausible Motion Forecasting
Keyword: diffusion
GenEval: An Object-Focused Framework for Evaluating Text-to-Image Alignment
Reflection-Equivariant Diffusion for 3D Structure Determination from Isotopologue Rotational Spectra in Natural Abundance
On the Evaluation of Generative Models in Distributed Learning Tasks
Language Agents for Detecting Implicit Stereotypes in Text-to-image Models at Scale
Progressive3D: Progressively Local Editing for Text-to-3D Content Creation with Complex Semantic Prompts
De novo protein design using geometric vector field networks
To Generate or Not? Safety-Driven Unlearned Diffusion Models Are Still Easy To Generate Unsafe Images ... For Now
IRAD: Implicit Representation-driven Image Resampling against Adversarial Attacks
InfoDiffusion: Information Entropy Aware Diffusion Process for Non-Autoregressive Text Generation
Image Super-resolution Via Latent Diffusion: A Sampling-space Mixture Of Experts And Frequency-augmented Decoder Approach
Quality Diversity through Human Feedback
Keyword: adaptive
Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection
Value-Biased Maximum Likelihood Estimation for Model-based Reinforcement Learning in Discounted Linear MDPs
Differentially Private Data Generation with Missing Data
WaveAttack: Asymmetric Frequency Obfuscation-based Backdoor Attacks Against Deep Neural Networks
MISAR: A Multimodal Instructional System with Augmented Reality
Action-Quantized Offline Reinforcement Learning for Robotic Skill Learning
AdaVis: Adaptive and Explainable Visualization Recommendation for Tabular Data
Hierarchical accompanying and inhibiting patterns on the spatial arrangement of taxis' local hotspots
Accelerate Presolve in Large-Scale Linear Programming via Reinforcement Learning
Learning to Generate Parameters of ConvNets for Unseen Image Data
Stochastic Optimization for Non-convex Problem with Inexact Hessian Matrix, Gradient, and Function
Evolving Bitcoin Custody
Improving Generalization of Alignment with Human Preferences through Group Invariant Learning
SegmATRon: Embodied Adaptive Semantic Segmentation for Indoor Environment
Ordered Reliability Direct Error Pattern Testing Decoding Algorithm
SHARCS: Efficient Transformers through Routing with Dynamic Width Sub-networks
Keyword: quantization
A Comprehensive Survey on Vector Database: Storage and Retrieval Technique, Challenge
Action-Quantized Offline Reinforcement Learning for Robotic Skill Learning
VQ-NeRF: Neural Reflectance Decomposition and Editing with Vector Quantization