Abstract
A fairly reliable trend in deep reinforcement learning is that the performance scales with the number of parameters, provided a complimentary scaling in amount of training data. As the appetite for large models increases, it is imperative to address, sooner than later, the potential problem of running out of high-quality demonstrations. In this case, instead of collecting only new data via costly human demonstrations or risking a simulation-to-real transfer with uncertain effects, it would be beneficial to leverage vast amounts of readily-available low-quality data. Since classical control algorithms such as behavior cloning or temporal difference learning cannot be used on reward-free or action-free data out-of-the-box, this solution warrants novel training paradigms for continuous control. We propose a simple algorithm called Diffused Value Function (DVF), which learns a joint multi-step model of the environment-robot interaction dynamics using a diffusion model. This model can be efficiently learned from state sequences (i.e., without access to reward functions nor actions), and subsequently used to estimate the value of each action out-of-the-box. We show how DVF can be used to efficiently capture the state visitation measure for multiple controllers, and show promising qualitative and quantitative results on challenging robotics benchmarks.
Expressivity Enhancement with Efficient Quadratic Neurons for Convolutional Neural Networks
Authors: Chuangtao Chen, Grace Li Zhang, Xunzhao Yin, Cheng Zhuo, Ulf Schlichtmann, Bing Li
Abstract
Convolutional neural networks (CNNs) have been successfully applied in a range of fields such as image classification and object segmentation. To improve their expressivity, various techniques, such as novel CNN architectures, have been explored. However, the performance gain from such techniques tends to diminish. To address this challenge, many researchers have shifted their focus to increasing the non-linearity of neurons, the fundamental building blocks of neural networks, to enhance the network expressivity. Nevertheless, most of these approaches incur a large number of parameters and thus formidable computation cost inevitably, impairing their efficiency to be deployed in practice. In this work, an efficient quadratic neuron structure is proposed to preserve the non-linearity with only negligible parameter and computation cost overhead. The proposed quadratic neuron can maximize the utilization of second-order computation information to improve the network performance. The experimental results have demonstrated that the proposed quadratic neuron can achieve a higher accuracy and a better computation efficiency in classification tasks compared with both linear neurons and non-linear neurons from previous works.
Referring to Screen Texts with Voice Assistants
Authors: Shruti Bhargava, Anand Dhoot, Ing-Marie Jonsson, Hoang Long Nguyen, Alkesh Patel, Hong Yu, Vincent Renkens
Abstract
Voice assistants help users make phone calls, send messages, create events, navigate, and do a lot more. However, assistants have limited capacity to understand their users' context. In this work, we aim to take a step in this direction. Our work dives into a new experience for users to refer to phone numbers, addresses, email addresses, URLs, and dates on their phone screens. Our focus lies in reference understanding, which becomes particularly interesting when multiple similar texts are present on screen, similar to visual grounding. We collect a dataset and propose a lightweight general-purpose model for this novel experience. Due to the high cost of consuming pixels directly, our system is designed to rely on the extracted text from the UI. Our model is modular, thus offering flexibility, improved interpretability, and efficient runtime memory utilization.
Novel Regression and Least Square Support Vector Machine Learning Technique for Air Pollution Forecasting
Abstract
Air pollution is the origination of particulate matter, chemicals, or biological substances that brings pain to either humans or other living creatures or instigates discomfort to the natural habitat and the airspace. Hence, air pollution remains one of the paramount environmental issues as far as metropolitan cities are concerned. Several air pollution benchmarks are even said to have a negative influence on human health. Also, improper detection of air pollution benchmarks results in severe complications for humans and living creatures. To address this aspect, a novel technique called, Discretized Regression and Least Square Support Vector (DR-LSSV) based air pollution forecasting is proposed. The results indicate that the proposed DR-LSSV Technique can efficiently enhance air pollution forecasting performance and outperforms the conventional machine learning methods in terms of air pollution forecasting accuracy, air pollution forecasting time, and false positive rate.
A Holistic Approach to Unifying Automatic Concept Extraction and Concept Importance Estimation
Authors: Thomas Fel, Victor Boutin, Mazda Moayeri, Rémi Cadène, Louis Bethune, Léo andéol, Mathieu Chalvidal, Thomas Serre
Abstract
In recent years, concept-based approaches have emerged as some of the most promising explainability methods to help us interpret the decisions of Artificial Neural Networks (ANNs). These methods seek to discover intelligible visual 'concepts' buried within the complex patterns of ANN activations in two key steps: (1) concept extraction followed by (2) importance estimation. While these two steps are shared across methods, they all differ in their specific implementations. Here, we introduce a unifying theoretical framework that comprehensively defines and clarifies these two steps. This framework offers several advantages as it allows us: (i) to propose new evaluation metrics for comparing different concept extraction approaches; (ii) to leverage modern attribution methods and evaluation metrics to extend and systematically evaluate state-of-the-art concept-based approaches and importance estimation techniques; (iii) to derive theoretical guarantees regarding the optimality of such methods. We further leverage our framework to try to tackle a crucial question in explainability: how to efficiently identify clusters of data points that are classified based on a similar shared strategy. To illustrate these findings and to highlight the main strategies of a model, we introduce a visual representation called the strategic cluster graph. Finally, we present https://serre-lab.github.io/Lens, a dedicated website that offers a complete compilation of these visualizations for all classes of the ImageNet dataset.
Online Prototype Alignment for Few-shot Policy Transfer
Abstract
Domain adaptation in reinforcement learning (RL) mainly deals with the changes of observation when transferring the policy to a new environment. Many traditional approaches of domain adaptation in RL manage to learn a mapping function between the source and target domain in explicit or implicit ways. However, they typically require access to abundant data from the target domain. Besides, they often rely on visual clues to learn the mapping function and may fail when the source domain looks quite different from the target domain. To address these problems, we propose a novel framework Online Prototype Alignment (OPA) to learn the mapping function based on the functional similarity of elements and is able to achieve the few-shot policy transfer within only several episodes. The key insight of OPA is to introduce an exploration mechanism that can interact with the unseen elements of the target domain in an efficient and purposeful manner, and then connect them with the seen elements in the source domain according to their functionalities (instead of visual clues). Experimental results show that when the target domain looks visually different from the source domain, OPA can achieve better transfer performance even with much fewer samples from the target domain, outperforming prior methods.
A New Probabilistic Distance Metric With Application In Gaussian Mixture Reduction
Authors: Ahmad Sajedi, Yuri A. Lawryshyn, Konstantinos N. Plataniotis
Abstract
This paper presents a new distance metric to compare two continuous probability density functions. The main advantage of this metric is that, unlike other statistical measurements, it can provide an analytic, closed-form expression for a mixture of Gaussian distributions while satisfying all metric properties. These characteristics enable fast, stable, and efficient calculations, which are highly desirable in real-world signal processing applications. The application in mind is Gaussian Mixture Reduction (GMR), which is widely used in density estimation, recursive tracking, and belief propagation. To address this problem, we developed a novel algorithm dubbed the Optimization-based Greedy GMR (OGGMR), which employs our metric as a criterion to approximate a high-order Gaussian mixture with a lower order. Experimental results show that the OGGMR algorithm is significantly faster and more efficient than state-of-the-art GMR algorithms while retaining the geometric shape of the original mixture.
Exploratory analysis of a measurement scale of an information security management system
Authors: Rúsbel Domínguez Domínguez (1), Omar A. Flores Laguna (1), José A. Sánchez-Valdez (2) ((1) Universidad de Montemorelos, (2) Tecnológico Nacional de México)
Abstract
This research shows the analysis of multiple factors that inhibit the implementation of an Information Security Management System (ISMS). The research data were collected from 143 respondents from two universities in northeastern Mexico, in faculties of engineering in related areas. In this study, the Information Security Management System Measurement Instrument (IM-ISMS) was validated. A scale of 24 items was obtained, divided into four factors: organizational policies and regulations, privacy, integrity and authenticity. The results of this study agree with the results found by [10] in which they pre-sent a model that complies with ISO/IEC 27002:2013 controls and security and privacy criteria to improve the ISMS. [48], Mentioned that the implementation of controls based on ISO standards can meet the requirements for cybersecurity best practices.A scale of 24 items was obtained, divided into four factors: organizational policies and regulations, privacy, integrity and authenticity. This version of the instrument meets the criteria established for its validity (KMO, Bartlett's test of sphericity). An extraction was performed by the minimum residuals method, an oblique rotation was performed by the promax method, when performing the rotation 17 of the 24 items were grouped in the corresponding factor. The final reliability of the scale was calculated by the Omega coefficient, in all the dimensions the coefficients were greater than .70, therefore the re-liability of the instrument is good.
Set-based state estimation and fault diagnosis using constrained zonotopes and applications
Abstract
This doctoral thesis develops new methods for set-based state estimation and active fault diagnosis (AFD) of (i) nonlinear discrete-time systems, (ii) discrete-time nonlinear systems whose trajectories satisfy nonlinear equality constraints (called invariants), (iii) linear descriptor systems, and (iv) joint state and parameter estimation of nonlinear descriptor systems. Set-based estimation aims to compute tight enclosures of the possible system states in each time step subject to unknown-but-bounded uncertainties. To address this issue, the present doctoral thesis proposes new methods for efficiently propagating constrained zonotopes (CZs) through nonlinear mappings. Besides, this thesis improves the standard prediction-update framework for systems with invariants using new algorithms for refining CZs based on nonlinear constraints. In addition, this thesis introduces a new approach for set-based AFD of a class of nonlinear discrete-time systems. An affine parametrization of the reachable sets is obtained for the design of an optimal input for set-based AFD. In addition, this thesis presents new methods based on CZs for set-valued state estimation and AFD of linear descriptor systems. Linear static constraints on the state variables can be directly incorporated into CZs. Moreover, this thesis proposes a new representation for unbounded sets based on zonotopes, which allows to develop methods for state estimation and AFD also of unstable linear descriptor systems, without the knowledge of an enclosure of all the trajectories of the system. This thesis also develops a new method for set-based joint state and parameter estimation of nonlinear descriptor systems using CZs in a unified framework. Lastly, this manuscript applies the proposed set-based state estimation and AFD methods using CZs to unmanned aerial vehicles, water distribution networks, and a lithium-ion cell.
Robust Reinforcement Learning through Efficient Adversarial Herding
Authors: Juncheng Dong, Hao-Lun Hsu, Qitong Gao, Vahid Tarokh, Miroslav Pajic
Abstract
Although reinforcement learning (RL) is considered the gold standard for policy design, it may not always provide a robust solution in various scenarios. This can result in severe performance degradation when the environment is exposed to potential disturbances. Adversarial training using a two-player max-min game has been proven effective in enhancing the robustness of RL agents. In this work, we extend the two-player game by introducing an adversarial herd, which involves a group of adversaries, in order to address ($\textit{i}$) the difficulty of the inner optimization problem, and ($\textit{ii}$) the potential over pessimism caused by the selection of a candidate adversary set that may include unlikely scenarios. We first prove that adversarial herds can efficiently approximate the inner optimization problem. Then we address the second issue by replacing the worst-case performance in the inner optimization with the average performance over the worst-$k$ adversaries. We evaluate the proposed method on multiple MuJoCo environments. Experimental results demonstrate that our approach consistently generates more robust policies.
Instant Multi-View Head Capture through Learnable Registration
Authors: Timo Bolkart, Tianye Li, Michael J. Black
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Existing methods for capturing datasets of 3D heads in dense semantic correspondence are slow, and commonly address the problem in two separate steps; multi-view stereo (MVS) reconstruction followed by non-rigid registration. To simplify this process, we introduce TEMPEH (Towards Estimation of 3D Meshes from Performances of Expressive Heads) to directly infer 3D heads in dense correspondence from calibrated multi-view images. Registering datasets of 3D scans typically requires manual parameter tuning to find the right balance between accurately fitting the scans surfaces and being robust to scanning noise and outliers. Instead, we propose to jointly register a 3D head dataset while training TEMPEH. Specifically, during training we minimize a geometric loss commonly used for surface registration, effectively leveraging TEMPEH as a regularizer. Our multi-view head inference builds on a volumetric feature representation that samples and fuses features from each view using camera calibration information. To account for partial occlusions and a large capture volume that enables head movements, we use view- and surface-aware feature fusion, and a spatial transformer-based head localization module, respectively. We use raw MVS scans as supervision during training, but, once trained, TEMPEH directly predicts 3D heads in dense correspondence without requiring scans. Predicting one head takes about 0.3 seconds with a median reconstruction error of 0.26 mm, 64% lower than the current state-of-the-art. This enables the efficient capture of large datasets containing multiple people and diverse facial motions. Code, model, and data are publicly available at https://tempeh.is.tue.mpg.de.
Constructing Printable Surfaces with View-Dependent Appearance
Abstract
We present a method for the digital fabrication of surfaces whose appearance varies based on viewing direction. The surfaces are constructed from a mesh of bars arranged in a self-occluding colored heightfield that creates the desired view-dependent effects. At the heart of our method is a novel and simple differentiable rendering algorithm specifically designed to render colored 3D heightfields and enable efficient calculation of the gradient of appearance with respect to heights and colors. This algorithm forms the basis of a coarse-to-fine ML-based optimization process that adjusts the heights and colors of the strips to minimize the loss between the desired and real surface appearance from each viewpoint, deriving meshes that can then be fabricated using a 3D printer. Using our method, we demonstrate both synthetic and real-world fabricated results with view-dependent appearance.
ELF Codes: Concatenated Codes with an Expurgating Linear Function as the Outer Code
Abstract
An expurgating linear function (ELF) is a linear outer code that disallows the low-weight codewords of the inner code. ELFs can be designed either to maximize the minimum distance or to minimize the codeword error rate (CER) of the expurgated code. List decoding of the inner code from the noiseless all-zeros codeword is an efficient way to identify ELFs that maximize the minimum distance of the expurgated code. For convolutional inner codes, this paper provides distance spectrum union (DSU) upper bounds on the CER of the concatenated code. For short codeword lengths, ELFs transform a good inner code into a great concatenated code. For a constant message size of $K=64$ bits or constant codeword blocklength of $N=152$ bits, an ELF can reduce the gap at CER $10^{-6}$ between the DSU and the random-coding union (RCU) bounds from over 1 dB for the inner code alone to 0.23 dB for the concatenated code. The DSU bounds can also characterize puncturing that mitigates the rate overhead of the ELF while maintaining the DSU-to-RCU gap. The reduction in DSU-to-RCU gap comes with a minimal increase in average complexity. List Viterbi decoding guided by the ELF approaches maximum likelihood (ML) decoding of the concatenated code, and average list size converges to 1 as SNR increases. Thus, average complexity is similar to Viterbi decoding on the trellis of the inner code. For rare large-magnitude noise events, which occur less often than the FER of the inner code, a deep search in the list finds the ML codeword.
Multi-objective Molecular Optimization for Opioid Use Disorder Treatment Using Generative Network Complex
Abstract
Opioid Use Disorder (OUD) has emerged as a significant global public health issue, with complex multifaceted conditions. Due to the lack of effective treatment options for various conditions, there is a pressing need for the discovery of new medications. In this study, we propose a deep generative model that combines a stochastic differential equation (SDE)-based diffusion modeling with the latent space of a pretrained autoencoder model. The molecular generator enables efficient generation of molecules that are effective on multiple targets, specifically the mu, kappa, and delta opioid receptors. Furthermore, we assess the ADMET (absorption, distribution, metabolism, excretion, and toxicity) properties of the generated molecules to identify drug-like compounds. To enhance the pharmacokinetic properties of some lead compounds, we employ a molecular optimization approach. We obtain a diverse set of drug-like molecules. We construct binding affinity predictors by integrating molecular fingerprints derived from autoencoder embeddings, transformer embeddings, and topological Laplacians with advanced machine learning algorithms. Further experimental studies are needed to evaluate the pharmacological effects of these drug-like compounds for OUD treatment. Our machine learning platform serves as a valuable tool in designing and optimizing effective molecules for addressing OUD.
Improving Opinion-based Question Answering Systems Through Label Error Detection and Overwrite
Authors: Xiao Yang, Ahmed K. Mohamed, Shashank Jain, Stanislav Peshterliev, Debojeet Chatterjee, Hanwen Zha, Nikita Bhalla, Gagan Aneja, Pranab Mohanty
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Abstract
Label error is a ubiquitous problem in annotated data. Large amounts of label error substantially degrades the quality of deep learning models. Existing methods to tackle the label error problem largely focus on the classification task, and either rely on task specific architecture or require non-trivial additional computations, which is undesirable or even unattainable for industry usage. In this paper, we propose LEDO: a model-agnostic and computationally efficient framework for Label Error Detection and Overwrite. LEDO is based on Monte Carlo Dropout combined with uncertainty metrics, and can be easily generalized to multiple tasks and data sets. Applying LEDO to an industry opinion-based question answering system demonstrates it is effective at improving accuracy in all the core models. Specifically, LEDO brings 1.1% MRR gain for the retrieval model, 1.5% PR AUC improvement for the machine reading comprehension model, and 0.9% rise in the Average Precision for the ranker, on top of the strong baselines with a large-scale social media dataset. Importantly, LEDO is computationally efficient compared to methods that require loss function change, and cost-effective as the resulting data can be used in the same continuous training pipeline for production. Further analysis shows that these gains come from an improved decision boundary after cleaning the label errors existed in the training data.
PaVa: a novel Path-based Valley-seeking clustering algorithm
Authors: Lin Ma, Conan Liu, Tiefeng Ma, Shuangzhe Liu
Abstract
Clustering methods are being applied to a wider range of scenarios involving more complex datasets, where the shapes of clusters tend to be arbitrary. In this paper, we propose a novel Path-based Valley-seeking clustering algorithm for arbitrarily shaped clusters. This work aims to seek the valleys among clusters and then individually extract clusters. Three vital techniques are used in this algorithm. First, path distance (minmax distance) is employed to transform the irregular boundaries among clusters, that is density valleys, into perfect spherical shells. Second, a suitable density measurement, $k$-distance, is employed to make adjustment on Minimum Spanning Tree, by which a robust minmax distance is calculated. Third, we seek the transformed density valleys by determining their centers and radius. First, the clusters are wrapped in spherical shells after the distance transformation, making the extraction process efficient even with clusters of arbitrary shape. Second, adjusted Minimum Spanning Tree enhances the robustness of minmax distance under different kinds of noise. Last, the number of clusters does not need to be inputted or decided manually due to the individual extraction process. After applying the proposed algorithm to several commonly used synthetic datasets, the results indicate that the Path-based Valley-seeking algorithm is accurate and efficient. The algorithm is based on the dissimilarity of objects, so it can be applied to a wide range of fields. Its performance on real-world datasets illustrates its versatility.
Energy Efficient RAN Slicing and Beams Selection for Multiplexing of Heterogeneous Services in 5G mmWave Networks
Authors: PraveenKumar Korrai, Eva Lagunas, Shree Krishna Sharma, Symeon Chatzinotas
Subjects: Information Theory (cs.IT); Signal Processing (eess.SP)
Abstract
In this paper, we study a RAN resource-slicing problem for energy-efficient communication in an orthogonal frequency division multiple access (OFDMA) based millimeter-wave (mmWave) downlink (DL) network consisting of enhanced mobile broadband (eMBB) and ultra-reliable low-latency communication (URLLC) services. Specifically, assuming a fixed set of predefined beams, we address an energy efficiency (EE) maximization problem to obtain the optimal beam selection, Resource Block (RB), and transmit power allocation policy to serve URLLC and eMBB users on the same physical radio resources. The problem is formulated as a mixed-integer non-linear fractional programming (MINLFP) problem considering minimum data rate and latency in packet delivery constraints. By leveraging the properties of fractional programming theory, we first transform the formulated non-convex optimization problem in fractional form into a tractable subtractive form. Subsequently, we solve the transformed problem using a two-loop iterative algorithm. The main resource-slicing problem is solved in the inner loop utilizing the difference of convex (DC) programming and successive convex approximation (SCA) techniques. Subsequently, the outer loop is solved using the Dinkelbach method to acquire an improved solution in every iteration until it converges. Our simulation results illustrate the performance gains of the proposed methodology with respect to baseline algorithms with the fixed and mixed resource grid models.
Using Collision Momentum in Deep Reinforcement Learning Based Adversarial Pedestrian Modeling
Authors: Dianwei Chen, Ekim Yurtsever, Keith Redmill, Umit Ozguner
Abstract
Recent research in pedestrian simulation often aims to develop realistic behaviors in various situations, but it is challenging for existing algorithms to generate behaviors that identify weaknesses in automated vehicles' performance in extreme and unlikely scenarios and edge cases. To address this, specialized pedestrian behavior algorithms are needed. Current research focuses on realistic trajectories using social force models and reinforcement learning based models. However, we propose a reinforcement learning algorithm that specifically targets collisions and better uncovers unique failure modes of automated vehicle controllers. Our algorithm is efficient and generates more severe collisions, allowing for the identification and correction of weaknesses in autonomous driving algorithms in complex and varied scenarios.
Hybrid and Oriented Harmonic Potentials for Safe Task Execution in Unknown Environment
Abstract
Harmonic potentials provide globally convergent potential fields that are provably free of local minima. Due to its analytical format, it is particularly suitable for generating safe and reliable robot navigation policies. However, for complex environments that consist of a large number of overlapping non-sphere obstacles, the computation of associated transformation functions can be tedious. This becomes more apparent when: (i) the workspace is initially unknown and the underlying potential fields are updated constantly as the robot explores it; (ii) the high-level mission consists of sequential navigation tasks among numerous regions, requiring the robot to switch between different potentials. Thus, this work proposes an efficient and automated scheme to construct harmonic potentials incrementally online as guided by the task automaton. A novel two-layer harmonic tree (HT) structure is introduced that facilitates the hybrid combination of oriented search algorithms for task planning and harmonic-based navigation controllers for non-holonomic robots. Both layers are adapted efficiently and jointly during online execution to reflect the actual feasibility and cost of navigation within the updated workspace. Global safety and convergence are ensured both for the high-level task plan and the low-level robot trajectory. Known issues such as oscillation or long-detours for purely potential-based methods and sharp-turns or high computation complexity for purely search-based methods are prevented. Extensive numerical simulation and hardware experiments are conducted against several strong baselines.
A Simple Unified Uncertainty-Guided Framework for Offline-to-Online Reinforcement Learning
Abstract
Offline reinforcement learning (RL) provides a promising solution to learning an agent fully relying on a data-driven paradigm. However, constrained by the limited quality of the offline dataset, its performance is often sub-optimal. Therefore, it is desired to further finetune the agent via extra online interactions before deployment. Unfortunately, offline-to-online RL can be challenging due to two main challenges: constrained exploratory behavior and state-action distribution shift. To this end, we propose a Simple Unified uNcertainty-Guided (SUNG) framework, which naturally unifies the solution to both challenges with the tool of uncertainty. Specifically, SUNG quantifies uncertainty via a VAE-based state-action visitation density estimator. To facilitate efficient exploration, SUNG presents a practical optimistic exploration strategy to select informative actions with both high value and high uncertainty. Moreover, SUNG develops an adaptive exploitation method by applying conservative offline RL objectives to high-uncertainty samples and standard online RL objectives to low-uncertainty samples to smoothly bridge offline and online stages. SUNG achieves state-of-the-art online finetuning performance when combined with different offline RL methods, across various environments and datasets in D4RL benchmark.
HAUSER: Towards Holistic and Automatic Evaluation of Simile Generation
Abstract
Similes play an imperative role in creative writing such as story and dialogue generation. Proper evaluation metrics are like a beacon guiding the research of simile generation (SG). However, it remains under-explored as to what criteria should be considered, how to quantify each criterion into metrics, and whether the metrics are effective for comprehensive, efficient, and reliable SG evaluation. To address the issues, we establish HAUSER, a holistic and automatic evaluation system for the SG task, which consists of five criteria from three perspectives and automatic metrics for each criterion. Through extensive experiments, we verify that our metrics are significantly more correlated with human ratings from each perspective compared with prior automatic metrics.
Ethical Aspects of ChatGPT in Software Engineering Research
Authors: Muhammad Azeem Akbar, Arif Ali Khan, Peng Liang
Abstract
ChatGPT can improve Software Engineering (SE) research practices by offering efficient, accessible information analysis and synthesis based on natural language interactions. However, ChatGPT could bring ethical challenges, encompassing plagiarism, privacy, data security, and the risk of generating biased or potentially detrimental data. This research aims to fill the given gap by elaborating on the key elements: motivators, demotivators, and ethical principles of using ChatGPT in SE research. To achieve this objective, we conducted a literature survey, identified the mentioned elements, and presented their relationships by developing a taxonomy. Further, the identified literature-based elements (motivators, demotivators, and ethical principles) were empirically evaluated by conducting a comprehensive questionnaire-based survey involving SE researchers. Additionally, we employed Interpretive Structure Modeling (ISM) approach to analyze the relationships between the ethical principles of using ChatGPT in SE research and develop a level based decision model. We further conducted a Cross-Impact Matrix Multiplication Applied to Classification (MICMAC) analysis to create a cluster-based decision model. These models aim to help SE researchers devise effective strategies for ethically integrating ChatGPT into SE research by following the identified principles through adopting the motivators and addressing the demotivators. The findings of this study will establish a benchmark for incorporating ChatGPT services in SE research with an emphasis on ethical considerations.
Action Recognition with Multi-stream Motion Modeling and Mutual Information Maximization
Abstract
Action recognition has long been a fundamental and intriguing problem in artificial intelligence. The task is challenging due to the high dimensionality nature of an action, as well as the subtle motion details to be considered. Current state-of-the-art approaches typically learn from articulated motion sequences in the straightforward 3D Euclidean space. However, the vanilla Euclidean space is not efficient for modeling important motion characteristics such as the joint-wise angular acceleration, which reveals the driving force behind the motion. Moreover, current methods typically attend to each channel equally and lack theoretical constrains on extracting task-relevant features from the input. In this paper, we seek to tackle these challenges from three aspects: (1) We propose to incorporate an acceleration representation, explicitly modeling the higher-order variations in motion. (2) We introduce a novel Stream-GCN network equipped with multi-stream components and channel attention, where different representations (i.e., streams) supplement each other towards a more precise action recognition while attention capitalizes on those important channels. (3) We explore feature-level supervision for maximizing the extraction of task-relevant information and formulate this into a mutual information loss. Empirically, our approach sets the new state-of-the-art performance on three benchmark datasets, NTU RGB+D, NTU RGB+D 120, and NW-UCLA. Our code is anonymously released at https://github.com/ActionR-Group/Stream-GCN, hoping to inspire the community.
Parametric Implicit Face Representation for Audio-Driven Facial Reenactment
Authors: Ricong Huang, Peiwen Lai, Yipeng Qin, Guanbin Li
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Audio-driven facial reenactment is a crucial technique that has a range of applications in film-making, virtual avatars and video conferences. Existing works either employ explicit intermediate face representations (e.g., 2D facial landmarks or 3D face models) or implicit ones (e.g., Neural Radiance Fields), thus suffering from the trade-offs between interpretability and expressive power, hence between controllability and quality of the results. In this work, we break these trade-offs with our novel parametric implicit face representation and propose a novel audio-driven facial reenactment framework that is both controllable and can generate high-quality talking heads. Specifically, our parametric implicit representation parameterizes the implicit representation with interpretable parameters of 3D face models, thereby taking the best of both explicit and implicit methods. In addition, we propose several new techniques to improve the three components of our framework, including i) incorporating contextual information into the audio-to-expression parameters encoding; ii) using conditional image synthesis to parameterize the implicit representation and implementing it with an innovative tri-plane structure for efficient learning; iii) formulating facial reenactment as a conditional image inpainting problem and proposing a novel data augmentation technique to improve model generalizability. Extensive experiments demonstrate that our method can generate more realistic results than previous methods with greater fidelity to the identities and talking styles of speakers.
Binary Radiance Fields
Authors: Seungjoo Shin, Jaesik Park
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
In this paper, we propose binary radiance fields (BiRF), a storage-efficient radiance field representation employing binary feature encoding that encodes local features using binary encoding parameters in a format of either $+1$ or $-1$. This binarization strategy lets us represent the feature grid with highly compact feature encoding and a dramatic reduction in storage size. Furthermore, our 2D-3D hybrid feature grid design enhances the compactness of feature encoding as the 3D grid includes main components while 2D grids capture details. In our experiments, binary radiance field representation successfully outperforms the reconstruction performance of state-of-the-art (SOTA) efficient radiance field models with lower storage allocation. In particular, our model achieves impressive results in static scene reconstruction, with a PSNR of 31.53 dB for Synthetic-NeRF scenes, 34.26 dB for Synthetic-NSVF scenes, 28.02 dB for Tanks and Temples scenes while only utilizing 0.7 MB, 0.8 MB, and 0.8 MB of storage space, respectively. We hope the proposed binary radiance field representation will make radiance fields more accessible without a storage bottleneck.
Invertible Bloom Lookup Tables with Less Memory and Less Randomness
Authors: Nils Fleischhacker, Kasper Green Larsen, Maciej Obremski, Mark Simkin
Subjects: Data Structures and Algorithms (cs.DS); Cryptography and Security (cs.CR)
Abstract
In this work we study Invertible Bloom Lookup Tables (IBLTs) with small failure probabilities. IBLTs are highly versatile data structures that have found applications in set reconciliation protocols, error-correcting codes, and even the design of advanced cryptographic primitives. For storing $n$ elements and ensuring correctness with probability at least $1 - \delta$, existing IBLT constructions require $\Omega(n(\frac{\log(1/\delta)}{\log(n)}+1))$ space and they crucially rely on fully random hash functions. We present new constructions of IBLTs that are simultaneously more space efficient and require less randomness. For storing $n$ elements with a failure probability of at most $\delta$, our data structure only requires $\mathcal{O}(n + \log(1/\delta)\log\log(1/\delta))$ space and $\mathcal{O}(\log(\log(n)/\delta))$-wise independent hash functions. As a key technical ingredient we show that hashing $n$ keys with any $k$-wise independent hash function $h:U \to [Cn]$ for some sufficiently large constant $C$ guarantees with probability $1 - 2^{-\Omega(k)}$ that at least $n/2$ keys will have a unique hash value. Proving this is highly non-trivial as $k$ approaches $n$. We believe that the techniques used to prove this statement may be of independent interest.
Practice with Graph-based ANN Algorithms on Sparse Data: Chi-square Two-tower model, HNSW, Sign Cauchy Projections
Abstract
Sparse data are common. The traditional handcrafted'' features are often sparse. Embedding vectors from trained models can also be very sparse, for example, embeddings trained via theReLu'' activation function. In this paper, we report our exploration of efficient search in sparse data with graph-based ANN algorithms (e.g., HNSW, or SONG which is the GPU version of HNSW), which are popular in industrial practice, e.g., search and ads (advertising). We experiment with the proprietary ads targeting application, as well as benchmark public datasets. For ads targeting, we train embeddings with the standard cosine two-tower'' model and we also develop thechi-square two-tower'' model. Both models produce (highly) sparse embeddings when they are integrated with the ReLu'' activation function. In EBR (embedding-based retrieval) applications, after we the embeddings are trained, the next crucial task is the approximate near neighbor (ANN) search for serving. While there are many ANN algorithms we can choose from, in this study, we focus on the graph-based ANN algorithm (e.g., HNSW-type). Sparse embeddings should help improve the efficiency of EBR. One benefit is the reduced memory cost for the embeddings. The other obvious benefit is the reduced computational time for evaluating similarities, because, for graph-based ANN algorithms such as HNSW, computing similarities is often the dominating cost. In addition to the effort on leveraging data sparsity for storage and computation, we also integratesign cauchy random projections'' (SignCRP) to hash vectors to bits, to further reduce the memory cost and speed up the ANN search. In NIPS'13, SignCRP was proposed to hash the chi-square similarity, which is a well-adopted nonlinear kernel in NLP and computer vision. Therefore, the chi-square two-tower model, SignCRP, and HNSW are now tightly integrated.
UOD: Universal One-shot Detection of Anatomical Landmarks
Abstract
One-shot medical landmark detection gains much attention and achieves great success for its label-efficient training process. However, existing one-shot learning methods are highly specialized in a single domain and suffer domain preference heavily in the situation of multi-domain unlabeled data. Moreover, one-shot learning is not robust that it faces performance drop when annotating a sub-optimal image. To tackle these issues, we resort to developing a domain-adaptive one-shot landmark detection framework for handling multi-domain medical images, named Universal One-shot Detection (UOD). UOD consists of two stages and two corresponding universal models which are designed as combinations of domain-specific modules and domain-shared modules. In the first stage, a domain-adaptive convolution model is self-supervised learned to generate pseudo landmark labels. In the second stage, we design a domain-adaptive transformer to eliminate domain preference and build the global context for multi-domain data. Even though only one annotated sample from each domain is available for training, the domain-shared modules help UOD aggregate all one-shot samples to detect more robust and accurate landmarks. We investigated both qualitatively and quantitatively the proposed UOD on three widely-used public X-ray datasets in different anatomical domains (i.e., head, hand, chest) and obtained state-of-the-art performances in each domain.
SqueezeLLM: Dense-and-Sparse Quantization
Authors: Sehoon Kim, Coleman Hooper, Amir Gholami, Zhen Dong, Xiuyu Li, Sheng Shen, Michael W. Mahoney, Kurt Keutzer
Subjects: Computation and Language (cs.CL); Machine Learning (cs.LG)
Abstract
Generative Large Language Models (LLMs) have demonstrated remarkable results for a wide range of tasks. However, deploying these models for inference has been a significant challenge due to their unprecedented resource requirements. This has forced existing deployment frameworks to use multi-GPU inference pipelines, which are often complex and costly, or to use smaller and less performant models. In this work, we demonstrate that the main bottleneck for generative inference with LLMs is memory bandwidth, rather than compute, specifically for single batch inference. While quantization has emerged as a promising solution by representing model weights with reduced precision, previous efforts have often resulted in notable performance degradation. To address this, we introduce SqueezeLLM, a post-training quantization framework that not only enables lossless compression to ultra-low precisions of up to 3-bit, but also achieves higher quantization performance under the same memory constraint. Our framework incorporates two novel ideas: (i) sensitivity-based non-uniform quantization, which searches for the optimal bit precision assignment based on second-order information; and (ii) the Dense-and-Sparse decomposition that stores outliers and sensitive weight values in an efficient sparse format. When applied to the LLaMA models, our 3-bit quantization significantly reduces the perplexity gap from the FP16 baseline by up to 2.1x as compared to the state-of-the-art methods with the same memory requirement. Furthermore, when deployed on an A6000 GPU, our quantized models achieve up to 2.3x speedup compared to the baseline. Our code is open-sourced and available online.
Sea Ice Segmentation From SAR Data by Convolutional Transformer Networks
Authors: Nicolae-Catalin Ristea, Andrei Anghel, Mihai Datcu
Subjects: Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV)
Abstract
Sea ice is a crucial component of the Earth's climate system and is highly sensitive to changes in temperature and atmospheric conditions. Accurate and timely measurement of sea ice parameters is important for understanding and predicting the impacts of climate change. Nevertheless, the amount of satellite data acquired over ice areas is huge, making the subjective measurements ineffective. Therefore, automated algorithms must be used in order to fully exploit the continuous data feeds coming from satellites. In this paper, we present a novel approach for sea ice segmentation based on SAR satellite imagery using hybrid convolutional transformer (ConvTr) networks. We show that our approach outperforms classical convolutional networks, while being considerably more efficient than pure transformer models. ConvTr obtained a mean intersection over union (mIoU) of 63.68% on the AI4Arctic data set, assuming an inference time of 120ms for a 400 x 400 squared km product.
Automating Microservices Test Failure Analysis using Kubernetes Cluster Logs
Authors: Pawan Kumar Sarika, Deepika Badampudi, Sai Prashanth Josyula, Muhammad Usman
Abstract
Kubernetes is a free, open-source container orchestration system for deploying and managing Docker containers that host microservices. Kubernetes cluster logs help in determining the reason for the failure. However, as systems become more complex, identifying failure reasons manually becomes more difficult and time-consuming. This study aims to identify effective and efficient classification algorithms to automatically determine the failure reason. We compare five classification algorithms, Support Vector Machines, K-Nearest Neighbors, Random Forest, Gradient Boosting Classifier, and Multilayer Perceptron. Our results indicate that Random Forest produces good accuracy while requiring fewer computational resources than other algorithms.
Few-shot Multi-domain Knowledge Rearming for Context-aware Defence against Advanced Persistent Threats
Authors: Gaolei Li, Yuanyuan Zhao, Wenqi Wei, Yuchen Liu
Subjects: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
Abstract
Advanced persistent threats (APTs) have novel features such as multi-stage penetration, highly-tailored intention, and evasive tactics. APTs defense requires fusing multi-dimensional Cyber threat intelligence data to identify attack intentions and conducts efficient knowledge discovery strategies by data-driven machine learning to recognize entity relationships. However, data-driven machine learning lacks generalization ability on fresh or unknown samples, reducing the accuracy and practicality of the defense model. Besides, the private deployment of these APT defense models on heterogeneous environments and various network devices requires significant investment in context awareness (such as known attack entities, continuous network states, and current security strategies). In this paper, we propose a few-shot multi-domain knowledge rearming (FMKR) scheme for context-aware defense against APTs. By completing multiple small tasks that are generated from different network domains with meta-learning, the FMKR firstly trains a model with good discrimination and generalization ability for fresh and unknown APT attacks. In each FMKR task, both threat intelligence and local entities are fused into the support/query sets in meta-learning to identify possible attack stages. Secondly, to rearm current security strategies, an finetuning-based deployment mechanism is proposed to transfer learned knowledge into the student model, while minimizing the defense cost. Compared to multiple model replacement strategies, the FMKR provides a faster response to attack behaviors while consuming less scheduling cost. Based on the feedback from multiple real users of the Industrial Internet of Things (IIoT) over 2 months, we demonstrate that the proposed scheme can improve the defense satisfaction rate.
Abstract
Recently, there has been a growing trend toward feature-based approaches for Online Action Detection (OAD). However, these approaches have limitations due to their fixed backbone design, which ignores the potential capability of a trainable backbone. In this paper, we propose the first end-to-end OAD model, termed E2E-LOAD, designed to address the major challenge of OAD, namely, long-term understanding and efficient online reasoning. Specifically, our proposed approach adopts an initial spatial model that is shared by all frames and maintains a long sequence cache for inference at a low computational cost. We also advocate an asymmetric spatial-temporal model for long-form and short-form modeling effectively. Furthermore, we propose a novel and efficient inference mechanism that accelerates heavy spatial-temporal exploration. Extensive ablation studies and experiments demonstrate the effectiveness and efficiency of our proposed method. Notably, we achieve 17.3 (+12.6) FPS for end-to-end OAD with 72.4%~(+1.2%), 90.3%~(+0.7%), and 48.1%~(+26.0%) mAP on THMOUS14, TVSeries, and HDD, respectively, which is 3x faster than previous approaches. The source code will be made publicly available.
Just a Second -- Scheduling Thousands of Time-Triggered Streams in Large-Scale Networks
Authors: Heiko Geppert, Frank Dürr, Sukanya Bhowmik, Kurt Rothermel
Subjects: Networking and Internet Architecture (cs.NI)
Abstract
Deterministic real-time communication with bounded delay is an essential requirement for many safety-critical cyber-physical systems, and has received much attention from major standardization bodies such as IEEE and IETF. In particular, Ethernet technology has been extended by time-triggered scheduling mechanisms in standards like TTEthernet and Time-Sensitive Networking. Although the scheduling mechanisms have become part of standards, the traffic planning algorithms to create time-triggered schedules are still an open and challenging research question due to the problem's high complexity. In particular, so-called plug-and-produce scenarios require the ability to extend schedules on the fly within seconds. The need for scalable scheduling and routing algorithms is further supported by large-scale distributed real-time systems like smart energy grids with tight communication requirements. In this paper, we tackle this challenge by proposing two novel algorithms called Hierarchical Heuristic Scheduling (H2S) and Cost-Efficient Lazy Forwarding Scheduling (CELF) to calculate time-triggered schedules for TTEthernet. H2S and CELF are highly efficient and scalable, calculating schedules for more than 45,000 streams on random networks with 1,000 bridges as well as a realistic energy grid network within sub-seconds to seconds.
Kernelized Reinforcement Learning with Order Optimal Regret Bounds
Abstract
Reinforcement learning (RL) has shown empirical success in various real world settings with complex models and large state-action spaces. The existing analytical results, however, typically focus on settings with a small number of state-actions or simple models such as linearly modeled state-action value functions. To derive RL policies that efficiently handle large state-action spaces with more general value functions, some recent works have considered nonlinear function approximation using kernel ridge regression. We propose $\pi$-KRVI, an optimistic modification of least-squares value iteration, when the state-action value function is represented by an RKHS. We prove the first order-optimal regret guarantees under a general setting. Our results show a significant polynomial in the number of episodes improvement over the state of the art. In particular, with highly non-smooth kernels (such as Neural Tangent kernel or some Mat\'ern kernels) the existing results lead to trivial (superlinear in the number of episodes) regret bounds. We show a sublinear regret bound that is order optimal in the case of Mat\'ern kernels where a lower bound on regret is known.
NAVER LABS Europe's Multilingual Speech Translation Systems for the IWSLT 2023 Low-Resource Track
Authors: Edward Gow-Smith, Alexandre Berard, Marcely Zanon Boito, Ioan Calapodescu
Abstract
This paper presents NAVER LABS Europe's systems for Tamasheq-French and Quechua-Spanish speech translation in the IWSLT 2023 Low-Resource track. Our work attempts to maximize translation quality in low-resource settings using multilingual parameter-efficient solutions that leverage strong pre-trained models. Our primary submission for Tamasheq outperforms the previous state of the art by 7.5 BLEU points on the IWSLT 2022 test set, and achieves 23.6 BLEU on this year's test set, outperforming the second best participant by 7.7 points. For Quechua, we also rank first and achieve 17.7 BLEU, despite having only two hours of translation data. Finally, we show that our proposed multilingual architecture is also competitive for high-resource languages, outperforming the best unconstrained submission to the IWSLT 2021 Multilingual track, despite using much less training data and compute.
A Cloud-based Machine Learning Pipeline for the Efficient Extraction of Insights from Customer Reviews
Authors: Robert Lakatos, Gergo Bogacsovics, Balazs Harangi, Istvan Lakatos, Attila Tiba, Janos Toth, Marianna Szabo, Andras Hajdu
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
Abstract
The efficiency of natural language processing has improved dramatically with the advent of machine learning models, particularly neural network-based solutions. However, some tasks are still challenging, especially when considering specific domains. In this paper, we present a cloud-based system that can extract insights from customer reviews using machine learning methods integrated into a pipeline. For topic modeling, our composite model uses transformer-based neural networks designed for natural language processing, vector embedding-based keyword extraction, and clustering. The elements of our model have been integrated and further developed to meet better the requirements of efficient information extraction, topic modeling of the extracted information, and user needs. Furthermore, our system can achieve better results than this task's existing topic modeling and keyword extraction solutions. Our approach is validated and compared with other state-of-the-art methods using publicly available datasets for benchmarking.
Efficient GPU implementation of a class of array permutations
Authors: Mathis Bouverot-Dupuis, Mary Sheeran
Subjects: Distributed, Parallel, and Cluster Computing (cs.DC)
Abstract
Optimal usage of the memory system is a key element of fast GPU algorithms. Unfortunately many common algorithms fail in this regard despite exhibiting great regularity in memory access patterns. In this paper we propose efficient kernels to permute the elements of an array, which can be used to improve the access patterns of many algorithms. We handle a class of permutations known as Bit Matrix Multiply Complement (BMMC) permutations, for which we design kernels of speed comparable to that of a simple array copy. This is a first step towards implementing a set of array combinators based on these permutations.
GEmo-CLAP: Gender-Attribute-Enhanced Contrastive Language-Audio Pretraining for Speech Emotion Recognition
Authors: Yu Pan, Yanni Hu, Yuguang Yang, Jixun Yao, Wen Fei, Lei Ma, Heng Lu
Subjects: Computation and Language (cs.CL); Multimedia (cs.MM); Sound (cs.SD); Audio and Speech Processing (eess.AS)
Abstract
Contrastive Language-Audio Pretraining (CLAP) has recently exhibited impressive success in diverse fields. In this paper, we propose GEmo-CLAP, a kind of efficient gender-attribute-enhanced CLAP model for speech emotion recognition (SER). Specifically, we first build an effective emotion CLAP model termed Emo-CLAP for SER, utilizing various self-supervised learning based pre-trained models. Then, considering the importance of the gender attribute in speech emotion modeling, two GEmo-CLAP approaches are further proposed to integrate the emotion and gender information of speech signals, forming more reasonable objectives. Extensive experiments conducted on the IEMOCAP corpus demonstrate that our proposed two GEmo-CLAP approaches consistently outperform the baseline Emo-CLAP with different pre-trained models, while also achieving superior recognition performance compared with other state-of-the-art methods.
Expanding the Scope of DAWN: A Novel Version for Weighted Shortest Path Problem
Authors: Yelai Feng
Subjects: Data Structures and Algorithms (cs.DS); Distributed, Parallel, and Cluster Computing (cs.DC)
Abstract
The shortest path problem is a typical problem in graph theory with wide potential applications. The state-of-the-art single-source shortest paths algorithm on the weight graph is the $\Delta$-stepping algorithm, which can efficiently process weighted graphs in parallel. DAWN is an algorithm that addresses the shortest path problem on unweighted graphs, and we propose a weighted version that can handle graphs with weights edges, while maintaining the high scalability and parallelism features as DAWN. The novel version requires $O(\mu m)$ and $O(\mu \cdot E{wcc})$ times on the connected and unconnected graphs for SSSP problems, respectively. $E{wcc}$ denote the number of edges included in the largest weakly connected component, and $\mu$ is a constant denoting the average number of path transformations in the tasks. We tested the weighted version on the real graphs from Stanford Network Analysis Platform and SuiteSparse Matrix Collection, which outperformed the solution of $\Delta$-stepping algorithm from Gunrock, achieving a speedup of 43.163$\times$.
Robustly Learning a Single Neuron via Sharpness
Authors: Puqian Wang, Nikos Zarifis, Ilias Diakonikolas, Jelena Diakonikolas
Subjects: Machine Learning (cs.LG); Data Structures and Algorithms (cs.DS); Optimization and Control (math.OC); Statistics Theory (math.ST); Machine Learning (stat.ML)
Abstract
We study the problem of learning a single neuron with respect to the $L_2^2$-loss in the presence of adversarial label noise. We give an efficient algorithm that, for a broad family of activations including ReLUs, approximates the optimal $L_2^2$-error within a constant factor. Our algorithm applies under much milder distributional assumptions compared to prior work. The key ingredient enabling our results is a novel connection to local error bounds from optimization theory.
A dual number formulation to efficiently compute higher order directional derivatives
Authors: R. Peón-Escalante, K. B. Cantún-Avila, O. Carvente, A. Espinosa-Romero, F. Peñuñuri
Subjects: Numerical Analysis (math.NA); Computational Engineering, Finance, and Science (cs.CE)
Abstract
This contribution proposes a new formulation to efficiently compute directional derivatives of order one to fourth. The formulation is based on automatic differentiation implemented with dual numbers. Directional derivatives are particular cases of symmetric multilinear forms; therefore, using their symmetric properties and their coordinate representation, we implement functions to calculate mixed partial derivatives. Moreover, with directional derivatives, we deduce concise formulas for the velocity, acceleration, jerk, and jounce/snap vectors. The utility of our formulation is proved with three examples. The first example presents a comparison against the forward mode of finite differences to compute the fourth-order directional derivative of a scalar function. To this end, we have coded the finite differences method to calculate partial derivatives until the fourth order, to any order of approximation. The second example presents efficient computations of the velocity, acceleration, jerk, and jounce/snap. Finally, the third example is related to the computation of some partial derivatives. The implemented code of the proposed formulation and the finite differences method is proportioned as additional material to this article.
WebGLM: Towards An Efficient Web-Enhanced Question Answering System with Human Preferences
Authors: Xiao Liu, Hanyu Lai, Hao Yu, Yifan Xu, Aohan Zeng, Zhengxiao Du, Peng Zhang, Yuxiao Dong, Jie Tang
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
Abstract
We present WebGLM, a web-enhanced question-answering system based on the General Language Model (GLM). Its goal is to augment a pre-trained large language model (LLM) with web search and retrieval capabilities while being efficient for real-world deployments. To achieve this, we develop WebGLM with strategies for the LLM-augmented retriever, bootstrapped generator, and human preference-aware scorer. Specifically, we identify and address the limitations of WebGPT (OpenAI), through which WebGLM is enabled with accuracy, efficiency, and cost-effectiveness advantages. In addition, we propose systematic criteria for evaluating web-enhanced QA systems. We conduct multi-dimensional human evaluation and quantitative ablation studies, which suggest the outperformance of the proposed WebGLM designs over existing systems. WebGLM with the 10-billion-parameter GLM (10B) is shown to perform better than the similar-sized WebGPT (13B) and even comparably to WebGPT (175B) in human evaluation. The code, demo, and data are at \url{https://github.com/THUDM/WebGLM}.
Continuous Cost Aggregation for Dual-Pixel Disparity Extraction
Abstract
Recent works have shown that depth information can be obtained from Dual-Pixel (DP) sensors. A DP arrangement provides two views in a single shot, thus resembling a stereo image pair with a tiny baseline. However, the different point spread function (PSF) per view, as well as the small disparity range, makes the use of typical stereo matching algorithms problematic. To address the above shortcomings, we propose a Continuous Cost Aggregation (CCA) scheme within a semi-global matching framework that is able to provide accurate continuous disparities from DP images. The proposed algorithm fits parabolas to matching costs and aggregates parabola coefficients along image paths. The aggregation step is performed subject to a quadratic constraint that not only enforces the disparity smoothness but also maintains the quadratic form of the total costs. This gives rise to an inherently efficient disparity propagation scheme with a pixel-wise minimization in closed-form. Furthermore, the continuous form allows for a robust multi-scale aggregation that better compensates for the varying PSF. Experiments on DP data from both DSLR and phone cameras show that the proposed scheme attains state-of-the-art performance in DP disparity estimation.
Oracle-Efficient Pessimism: Offline Policy Optimization in Contextual Bandits
Authors: Lequn Wang, Akshay Krishnamurthy, Aleksandrs Slivkins
Abstract
We consider policy optimization in contextual bandits, where one is given a fixed dataset of logged interactions. While pessimistic regularizers are typically used to mitigate distribution shift, prior implementations thereof are not computationally efficient. We present the first oracle-efficient algorithm for pessimistic policy optimization: it reduces to supervised learning, leading to broad applicability. We also obtain best-effort statistical guarantees analogous to those for pessimistic approaches in prior work. We instantiate our approach for both discrete and continuous actions. We perform extensive experiments in both settings, showing advantage over unregularized policy optimization across a wide range of configurations.
GPT-Calls: Enhancing Call Segmentation and Tagging by Generating Synthetic Conversations via Large Language Models
Authors: Itzik Malkiel, Uri Alon, Yakir Yehuda, Shahar Keren, Oren Barkan, Royi Ronen, Noam Koenigstein
Subjects: Computation and Language (cs.CL); Machine Learning (cs.LG)
Abstract
Transcriptions of phone calls are of significant value across diverse fields, such as sales, customer service, healthcare, and law enforcement. Nevertheless, the analysis of these recorded conversations can be an arduous and time-intensive process, especially when dealing with extended or multifaceted dialogues. In this work, we propose a novel method, GPT-distilled Calls Segmentation and Tagging (GPT-Calls), for efficient and accurate call segmentation and topic extraction. GPT-Calls is composed of offline and online phases. The offline phase is applied once to a given list of topics and involves generating a distribution of synthetic sentences for each topic using a GPT model and extracting anchor vectors. The online phase is applied to every call separately and scores the similarity between the transcripted conversation and the topic anchors found in the offline phase. Then, time domain analysis is applied to the similarity scores to group utterances into segments and tag them with topics. The proposed paradigm provides an accurate and efficient method for call segmentation and topic extraction that does not require labeled data, thus making it a versatile approach applicable to various domains. Our algorithm operates in production under Dynamics 365 Sales Conversation Intelligence, and our research is based on real sales conversations gathered from various Dynamics 365 Sales tenants.
Parting with Misconceptions about Learning-based Vehicle Motion Planning
Authors: Daniel Dauner, Marcel Hallgarten, Andreas Geiger, Kashyap Chitta
Abstract
The release of nuPlan marks a new era in vehicle motion planning research, offering the first large-scale real-world dataset and evaluation schemes requiring both precise short-term planning and long-horizon ego-forecasting. Existing systems struggle to simultaneously meet both requirements. Indeed, we find that these tasks are fundamentally misaligned and should be addressed independently. We further assess the current state of closed-loop planning in the field, revealing the limitations of learning-based methods in complex real-world scenarios and the value of simple rule-based priors such as centerline selection through lane graph search algorithms. More surprisingly, for the open-loop sub-task, we observe that the best results are achieved when using only this centerline as scene context (\ie, ignoring all information regarding the map and other agents). Combining these insights, we propose an extremely simple and efficient planner which outperforms an extensive set of competitors, winning the nuPlan planning challenge 2023.
One-for-All: Generalized LoRA for Parameter-Efficient Fine-tuning
Abstract
We present Generalized LoRA (GLoRA), an advanced approach for universal parameter-efficient fine-tuning tasks. Enhancing Low-Rank Adaptation (LoRA), GLoRA employs a generalized prompt module to optimize pre-trained model weights and adjust intermediate activations, providing more flexibility and capability across diverse tasks and datasets. Moreover, GLoRA facilitates efficient parameter adaptation by employing a scalable, modular, layer-wise structure search that learns individual adapter of each layer. Originating from a unified mathematical formulation, GLoRA exhibits strong transfer learning, few-shot learning and domain generalization abilities, as it adjusts to new tasks through additional dimensions on weights and activations. Comprehensive experiments demonstrate that GLoRA outperforms all previous methods in natural, specialized, and structured benchmarks, achieving superior accuracy with fewer parameters and computations on various datasets. Furthermore, our structural re-parameterization design ensures that GLoRA incurs no extra inference cost, rendering it a practical solution for resource-limited applications. Code is available at: https://github.com/Arnav0400/ViT-Slim/tree/master/GLoRA.
Keyword: faster
A New Probabilistic Distance Metric With Application In Gaussian Mixture Reduction
Authors: Ahmad Sajedi, Yuri A. Lawryshyn, Konstantinos N. Plataniotis
Abstract
This paper presents a new distance metric to compare two continuous probability density functions. The main advantage of this metric is that, unlike other statistical measurements, it can provide an analytic, closed-form expression for a mixture of Gaussian distributions while satisfying all metric properties. These characteristics enable fast, stable, and efficient calculations, which are highly desirable in real-world signal processing applications. The application in mind is Gaussian Mixture Reduction (GMR), which is widely used in density estimation, recursive tracking, and belief propagation. To address this problem, we developed a novel algorithm dubbed the Optimization-based Greedy GMR (OGGMR), which employs our metric as a criterion to approximate a high-order Gaussian mixture with a lower order. Experimental results show that the OGGMR algorithm is significantly faster and more efficient than state-of-the-art GMR algorithms while retaining the geometric shape of the original mixture.
FIRE: An Optimization Approach for Fast Interpretable Rule Extraction
Abstract
We present FIRE, Fast Interpretable Rule Extraction, an optimization-based framework to extract a small but useful collection of decision rules from tree ensembles. FIRE selects sparse representative subsets of rules from tree ensembles, that are easy for a practitioner to examine. To further enhance the interpretability of the extracted model, FIRE encourages fusing rules during selection, so that many of the selected decision rules share common antecedents. The optimization framework utilizes a fusion regularization penalty to accomplish this, along with a non-convex sparsity-inducing penalty to aggressively select rules. Optimization problems in FIRE pose a challenge to off-the-shelf solvers due to problem scale and the non-convexity of the penalties. To address this, making use of problem-structure, we develop a specialized solver based on block coordinate descent principles; our solver performs up to 40x faster than existing solvers. We show in our experiments that FIRE outperforms state-of-the-art rule ensemble algorithms at building sparse rule sets, and can deliver more interpretable models compared to existing methods.
3D molecule generation by denoising voxel grids
Authors: Pedro O. Pinheiro, Joshua Rackers, Joseph Kleinhenz, Michael Maser, Omar Mahmood, Andrew Martin Watkins, Stephen Ra, Vishnu Sresht, Saeed Saremi
Abstract
We propose a new score-based approach to generate 3D molecules represented as atomic densities on regular grids. First, we train a denoising neural network that learns to map from a smooth distribution of noisy molecules to the distribution of real molecules. Then, we follow the neural empirical Bayes framework [Saremi and Hyvarinen, 2019] and generate molecules in two steps: (i) sample noisy density grids from a smooth distribution via underdamped Langevin Markov chain Monte Carlo, and (ii) recover the ``clean'' molecule by denoising the noisy grid with a single step. Our method, VoxMol, generates molecules in a fundamentally different way than the current state of the art (i.e., diffusion models applied to atom point clouds). It differs in terms of the data representation, the noise model, the network architecture and the generative modeling algorithm. VoxMol achieves comparable results to state of the art on unconditional 3D molecule generation while being simpler to train and faster to generate molecules.
Learning Unnormalized Statistical Models via Compositional Optimization
Abstract
Learning unnormalized statistical models (e.g., energy-based models) is computationally challenging due to the complexity of handling the partition function. To eschew this complexity, noise-contrastive estimation~(NCE) has been proposed by formulating the objective as the logistic loss of the real data and the artificial noise. However, as found in previous works, NCE may perform poorly in many tasks due to its flat loss landscape and slow convergence. In this paper, we study it a direct approach for optimizing the negative log-likelihood of unnormalized models from the perspective of compositional optimization. To tackle the partition function, a noise distribution is introduced such that the log partition function can be written as a compositional function whose inner function can be estimated with stochastic samples. Hence, the objective can be optimized by stochastic compositional optimization algorithms. Despite being a simple method, we demonstrate that it is more favorable than NCE by (1) establishing a fast convergence rate and quantifying its dependence on the noise distribution through the variance of stochastic estimators; (2) developing better results for one-dimensional Gaussian mean estimation by showing our objective has a much favorable loss landscape and hence our method enjoys faster convergence; (3) demonstrating better performance on multiple applications, including density estimation, out-of-distribution detection, and real image generation.
Galactic: Scaling End-to-End Reinforcement Learning for Rearrangement at 100k Steps-Per-Second
Authors: Vincent-Pierre Berges, Andrew Szot, Devendra Singh Chaplot, Aaron Gokaslan, Roozbeh Mottaghi, Dhruv Batra, Eric Undersander
Abstract
We present Galactic, a large-scale simulation and reinforcement-learning (RL) framework for robotic mobile manipulation in indoor environments. Specifically, a Fetch robot (equipped with a mobile base, 7DoF arm, RGBD camera, egomotion, and onboard sensing) is spawned in a home environment and asked to rearrange objects - by navigating to an object, picking it up, navigating to a target location, and then placing the object at the target location. Galactic is fast. In terms of simulation speed (rendering + physics), Galactic achieves over 421,000 steps-per-second (SPS) on an 8-GPU node, which is 54x faster than Habitat 2.0 (7699 SPS). More importantly, Galactic was designed to optimize the entire rendering + physics + RL interplay since any bottleneck in the interplay slows down training. In terms of simulation+RL speed (rendering + physics + inference + learning), Galactic achieves over 108,000 SPS, which 88x faster than Habitat 2.0 (1243 SPS). These massive speed-ups not only drastically cut the wall-clock training time of existing experiments, but also unlock an unprecedented scale of new experiments. First, Galactic can train a mobile pick skill to >80% accuracy in under 16 minutes, a 100x speedup compared to the over 24 hours it takes to train the same skill in Habitat 2.0. Second, we use Galactic to perform the largest-scale experiment to date for rearrangement using 5B steps of experience in 46 hours, which is equivalent to 20 years of robot experience. This scaling results in a single neural network composed of task-agnostic components achieving 85% success in GeometricGoal rearrangement, compared to 0% success reported in Habitat 2.0 for the same approach. The code is available at github.com/facebookresearch/galactic.
Few-shot Multi-domain Knowledge Rearming for Context-aware Defence against Advanced Persistent Threats
Authors: Gaolei Li, Yuanyuan Zhao, Wenqi Wei, Yuchen Liu
Subjects: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
Abstract
Advanced persistent threats (APTs) have novel features such as multi-stage penetration, highly-tailored intention, and evasive tactics. APTs defense requires fusing multi-dimensional Cyber threat intelligence data to identify attack intentions and conducts efficient knowledge discovery strategies by data-driven machine learning to recognize entity relationships. However, data-driven machine learning lacks generalization ability on fresh or unknown samples, reducing the accuracy and practicality of the defense model. Besides, the private deployment of these APT defense models on heterogeneous environments and various network devices requires significant investment in context awareness (such as known attack entities, continuous network states, and current security strategies). In this paper, we propose a few-shot multi-domain knowledge rearming (FMKR) scheme for context-aware defense against APTs. By completing multiple small tasks that are generated from different network domains with meta-learning, the FMKR firstly trains a model with good discrimination and generalization ability for fresh and unknown APT attacks. In each FMKR task, both threat intelligence and local entities are fused into the support/query sets in meta-learning to identify possible attack stages. Secondly, to rearm current security strategies, an finetuning-based deployment mechanism is proposed to transfer learned knowledge into the student model, while minimizing the defense cost. Compared to multiple model replacement strategies, the FMKR provides a faster response to attack behaviors while consuming less scheduling cost. Based on the feedback from multiple real users of the Industrial Internet of Things (IIoT) over 2 months, we demonstrate that the proposed scheme can improve the defense satisfaction rate.
Abstract
Recently, there has been a growing trend toward feature-based approaches for Online Action Detection (OAD). However, these approaches have limitations due to their fixed backbone design, which ignores the potential capability of a trainable backbone. In this paper, we propose the first end-to-end OAD model, termed E2E-LOAD, designed to address the major challenge of OAD, namely, long-term understanding and efficient online reasoning. Specifically, our proposed approach adopts an initial spatial model that is shared by all frames and maintains a long sequence cache for inference at a low computational cost. We also advocate an asymmetric spatial-temporal model for long-form and short-form modeling effectively. Furthermore, we propose a novel and efficient inference mechanism that accelerates heavy spatial-temporal exploration. Extensive ablation studies and experiments demonstrate the effectiveness and efficiency of our proposed method. Notably, we achieve 17.3 (+12.6) FPS for end-to-end OAD with 72.4%~(+1.2%), 90.3%~(+0.7%), and 48.1%~(+26.0%) mAP on THMOUS14, TVSeries, and HDD, respectively, which is 3x faster than previous approaches. The source code will be made publicly available.
DreamDecompiler: Improved Bayesian Program Learning by Decompiling Amortised Knowledge
Authors: Alessandro B. Palmarini, Christopher G. Lucas, N. Siddharth
Abstract
Solving program induction problems requires searching through an enormous space of possibilities. DreamCoder is an inductive program synthesis system that, whilst solving problems, learns to simplify search in an iterative wake-sleep procedure. The cost of search is amortised by training a neural search policy, reducing search breadth and effectively "compiling" useful information to compose program solutions across tasks. Additionally, a library of program components is learnt to express discovered solutions in fewer components, reducing search depth. In DreamCoder, the neural search policy has only an indirect effect on the library learnt through the program solutions it helps discover. We present an approach for library learning that directly leverages the neural search policy, effectively "decompiling" its amortised knowledge to extract relevant program components. This provides stronger amortised inference: the amortised knowledge learnt to reduce search breadth is now also used to reduce search depth. We integrate our approach with DreamCoder and demonstrate faster domain proficiency with improved generalisation on a range of domains, particularly when fewer example solutions are available.
CAMEO: A Causal Transfer Learning Approach for Performance Optimization of Configurable Computer Systems
Abstract
Modern computer systems are highly-configurable, with hundreds of configuration options interacting, resulting in enormous configuration space. As a result, optimizing performance goals (e.g., latency) in such systems is challenging. Worse, owing to evolving application requirements and user specifications, these systems face frequent uncertainties in their environments (e.g., hardware and workload change), making performance optimization even more challenging. Recently, transfer learning has been applied to address this problem by reusing knowledge from the offline configuration measurements of an old environment, aka, source to a new environment, aka, target. These approaches typically rely on predictive machine learning (ML) models to guide the search for finding interventions to optimize performance. However, previous empirical research showed that statistical models might perform poorly when the deployment environment changes because the independent and identically distributed (i.i.d.) assumption no longer holds. To address this issue, we propose Cameo -- a method that sidesteps these limitations by identifying invariant causal predictors under environmental changes, enabling the optimization process to operate on a reduced search space, leading to faster system performance optimization. We demonstrate significant performance improvements over the state-of-the-art optimization methods on five highly configurable computer systems, including three MLperf deep learning benchmark systems, a video analytics pipeline, and a database system, and studied the effectiveness in design explorations with different varieties and severity of environmental changes and show the scalability of our approach to colossal configuration spaces.
Keyword: mobile
Ethical Considerations of AR Applications in Smartphones; A Systematic Literature Review of Consumer Perspectives
Authors: Nicola J Wood
Subjects: Human-Computer Interaction (cs.HC); Computers and Society (cs.CY)
Abstract
This study focuses on the ethical considerations that a consumer perceives with augmented reality (AR) in the context of smartphone applications. Through a systematic review, this research can provide an understanding and ability for developers, product managers, digital marketers and associated business professionals to effectively implement and deploy mobile AR related applications and campaigns, with consideration to the perceptions of the ethical considerations that consumers have of this growing technology. The rise in digital transformation and new technologies paved this research agenda. Trends in the data revealed two overarching factors of 'Benefits' and 'Ethical Considerations'. Within these two factors, several consumer perceived themes were identified with regards to AR applications and their association categorised either positive, negative or neutral. 'Benefits' revealed 3 consistent themes of personalisation, interactivity and information acquisition. 'Ethical Considerations' revealed consistent patterns of educational awareness, privacy, transparency and security. From identifying the consumer perceptions, business professionals can strategically address and or challenge the inherent limitations and their associations during AR application development, product adoption strategies or marketing purposes.
Learning Any-View 6DoF Robotic Grasping in Cluttered Scenes via Neural Surface Rendering
Abstract
Robotic manipulation is critical for admitting robotic agents to various application domains, like intelligent assistance. A major challenge therein is the effective 6DoF grasping of objects in cluttered environments from any viewpoint without requiring additional scene exploration. We introduce $\textit{NeuGraspNet}$, a novel method for 6DoF grasp detection that leverages recent advances in neural volumetric representations and surface rendering. Our approach learns both global (scene-level) and local (grasp-level) neural surface representations, enabling effective and fully implicit 6DoF grasp quality prediction, even in unseen parts of the scene. Further, we reinterpret grasping as a local neural surface rendering problem, allowing the model to encode the interaction between the robot's end-effector and the object's surface geometry. NeuGraspNet operates on single viewpoints and can sample grasp candidates in occluded scenes, outperforming existing implicit and semi-implicit baseline methods in the literature. We demonstrate the real-world applicability of NeuGraspNet with a mobile manipulator robot, grasping in open spaces with clutter by rendering the scene, reasoning about graspable areas of different objects, and selecting grasps likely to succeed without colliding with the environment. Visit our project website: https://sites.google.com/view/neugraspnet
SoK: Decoding the Super App Enigma: The Security Mechanisms, Threats, and Trade-offs in OS-alike Apps
Authors: Yuqing Yang, Chao Wang, Yue Zhang, Zhiqiang Lin
Abstract
The super app paradigm, exemplified by platforms such as WeChat and AliPay, has revolutionized the mobile app landscape by enabling third-party developers to deploy add-ons within these apps. These add-ons, known as miniapps, leverage user data hosted by the super app platforms to provide a wide range of services, such as shopping and gaming. With the rise of miniapps, super apps have transformed into "operating systems", offering encapsulated APIs to miniapp developers as well as in-app miniapp stores for users to explore and download miniapps. In this paper, we provide the first systematic study to consolidate the current state of knowledge in this field from the security perspective: the security measures, threats, and trade-offs of this paradigm. Specifically, we summarize 13 security mechanisms and 10 security threats in super app platforms, followed by a root cause analysis revealing that the security assumptions still may be violated due to issues in underlying systems, implementation of isolation, and vetting. Additionally, we also systematize open problems and trade-offs that need to be addressed by future works to help enhance the security and privacy of this new paradigm.
Energy Efficient RAN Slicing and Beams Selection for Multiplexing of Heterogeneous Services in 5G mmWave Networks
Authors: PraveenKumar Korrai, Eva Lagunas, Shree Krishna Sharma, Symeon Chatzinotas
Subjects: Information Theory (cs.IT); Signal Processing (eess.SP)
Abstract
In this paper, we study a RAN resource-slicing problem for energy-efficient communication in an orthogonal frequency division multiple access (OFDMA) based millimeter-wave (mmWave) downlink (DL) network consisting of enhanced mobile broadband (eMBB) and ultra-reliable low-latency communication (URLLC) services. Specifically, assuming a fixed set of predefined beams, we address an energy efficiency (EE) maximization problem to obtain the optimal beam selection, Resource Block (RB), and transmit power allocation policy to serve URLLC and eMBB users on the same physical radio resources. The problem is formulated as a mixed-integer non-linear fractional programming (MINLFP) problem considering minimum data rate and latency in packet delivery constraints. By leveraging the properties of fractional programming theory, we first transform the formulated non-convex optimization problem in fractional form into a tractable subtractive form. Subsequently, we solve the transformed problem using a two-loop iterative algorithm. The main resource-slicing problem is solved in the inner loop utilizing the difference of convex (DC) programming and successive convex approximation (SCA) techniques. Subsequently, the outer loop is solved using the Dinkelbach method to acquire an improved solution in every iteration until it converges. Our simulation results illustrate the performance gains of the proposed methodology with respect to baseline algorithms with the fixed and mixed resource grid models.
Galactic: Scaling End-to-End Reinforcement Learning for Rearrangement at 100k Steps-Per-Second
Authors: Vincent-Pierre Berges, Andrew Szot, Devendra Singh Chaplot, Aaron Gokaslan, Roozbeh Mottaghi, Dhruv Batra, Eric Undersander
Abstract
We present Galactic, a large-scale simulation and reinforcement-learning (RL) framework for robotic mobile manipulation in indoor environments. Specifically, a Fetch robot (equipped with a mobile base, 7DoF arm, RGBD camera, egomotion, and onboard sensing) is spawned in a home environment and asked to rearrange objects - by navigating to an object, picking it up, navigating to a target location, and then placing the object at the target location. Galactic is fast. In terms of simulation speed (rendering + physics), Galactic achieves over 421,000 steps-per-second (SPS) on an 8-GPU node, which is 54x faster than Habitat 2.0 (7699 SPS). More importantly, Galactic was designed to optimize the entire rendering + physics + RL interplay since any bottleneck in the interplay slows down training. In terms of simulation+RL speed (rendering + physics + inference + learning), Galactic achieves over 108,000 SPS, which 88x faster than Habitat 2.0 (1243 SPS). These massive speed-ups not only drastically cut the wall-clock training time of existing experiments, but also unlock an unprecedented scale of new experiments. First, Galactic can train a mobile pick skill to >80% accuracy in under 16 minutes, a 100x speedup compared to the over 24 hours it takes to train the same skill in Habitat 2.0. Second, we use Galactic to perform the largest-scale experiment to date for rearrangement using 5B steps of experience in 46 hours, which is equivalent to 20 years of robot experience. This scaling results in a single neural network composed of task-agnostic components achieving 85% success in GeometricGoal rearrangement, compared to 0% success reported in Habitat 2.0 for the same approach. The code is available at github.com/facebookresearch/galactic.
Keyword: pruning
There is no result
Keyword: diffusion
Value function estimation using conditional diffusion models for control
Authors: Bogdan Mazoure, Walter Talbott, Miguel Angel Bautista, Devon Hjelm, Alexander Toshev, Josh Susskind
Abstract
A fairly reliable trend in deep reinforcement learning is that the performance scales with the number of parameters, provided a complimentary scaling in amount of training data. As the appetite for large models increases, it is imperative to address, sooner than later, the potential problem of running out of high-quality demonstrations. In this case, instead of collecting only new data via costly human demonstrations or risking a simulation-to-real transfer with uncertain effects, it would be beneficial to leverage vast amounts of readily-available low-quality data. Since classical control algorithms such as behavior cloning or temporal difference learning cannot be used on reward-free or action-free data out-of-the-box, this solution warrants novel training paradigms for continuous control. We propose a simple algorithm called Diffused Value Function (DVF), which learns a joint multi-step model of the environment-robot interaction dynamics using a diffusion model. This model can be efficiently learned from state sequences (i.e., without access to reward functions nor actions), and subsequently used to estimate the value of each action out-of-the-box. We show how DVF can be used to efficiently capture the state visitation measure for multiple controllers, and show promising qualitative and quantitative results on challenging robotics benchmarks.
Abstract
The diffusion maps embedding of data lying on a manifold have shown success in tasks ranging from dimensionality reduction and clustering, to data visualization. In this work, we consider embedding data sets which were sampled from a manifold which is closed under the action of a continuous matrix group. An example of such a data set are images who's planar rotations are arbitrary. The G-invariant graph Laplacian, introduced in a previous work of the authors, admits eigenfunctions in the form of tensor products between the elements of the irreducible unitary representations of the group and eigenvectors of certain matrices. We employ these eigenfunctions to derive diffusion maps that intrinsically account for the group action on the data. In particular, we construct both equivariant and invariant embeddings which can be used naturally to cluster and align the data points. We demonstrate the effectiveness of our construction with simulated data.
3D molecule generation by denoising voxel grids
Authors: Pedro O. Pinheiro, Joshua Rackers, Joseph Kleinhenz, Michael Maser, Omar Mahmood, Andrew Martin Watkins, Stephen Ra, Vishnu Sresht, Saeed Saremi
Abstract
We propose a new score-based approach to generate 3D molecules represented as atomic densities on regular grids. First, we train a denoising neural network that learns to map from a smooth distribution of noisy molecules to the distribution of real molecules. Then, we follow the neural empirical Bayes framework [Saremi and Hyvarinen, 2019] and generate molecules in two steps: (i) sample noisy density grids from a smooth distribution via underdamped Langevin Markov chain Monte Carlo, and (ii) recover the ``clean'' molecule by denoising the noisy grid with a single step. Our method, VoxMol, generates molecules in a fundamentally different way than the current state of the art (i.e., diffusion models applied to atom point clouds). It differs in terms of the data representation, the noise model, the network architecture and the generative modeling algorithm. VoxMol achieves comparable results to state of the art on unconditional 3D molecule generation while being simpler to train and faster to generate molecules.
Multi-objective Molecular Optimization for Opioid Use Disorder Treatment Using Generative Network Complex
Abstract
Opioid Use Disorder (OUD) has emerged as a significant global public health issue, with complex multifaceted conditions. Due to the lack of effective treatment options for various conditions, there is a pressing need for the discovery of new medications. In this study, we propose a deep generative model that combines a stochastic differential equation (SDE)-based diffusion modeling with the latent space of a pretrained autoencoder model. The molecular generator enables efficient generation of molecules that are effective on multiple targets, specifically the mu, kappa, and delta opioid receptors. Furthermore, we assess the ADMET (absorption, distribution, metabolism, excretion, and toxicity) properties of the generated molecules to identify drug-like compounds. To enhance the pharmacokinetic properties of some lead compounds, we employ a molecular optimization approach. We obtain a diverse set of drug-like molecules. We construct binding affinity predictors by integrating molecular fingerprints derived from autoencoder embeddings, transformer embeddings, and topological Laplacians with advanced machine learning algorithms. Further experimental studies are needed to evaluate the pharmacological effects of these drug-like compounds for OUD treatment. Our machine learning platform serves as a valuable tool in designing and optimizing effective molecules for addressing OUD.
User-defined Event Sampling and Uncertainty Quantification in Diffusion Models for Physical Dynamical Systems
Authors: Marc Finzi, Anudhyan Boral, Andrew Gordon Wilson, Fei Sha, Leonardo Zepeda-Núñez
Abstract
Diffusion models are a class of probabilistic generative models that have been widely used as a prior for image processing tasks like text conditional generation and inpainting. We demonstrate that these models can be adapted to make predictions and provide uncertainty quantification for chaotic dynamical systems. In these applications, diffusion models can implicitly represent knowledge about outliers and extreme events; however, querying that knowledge through conditional sampling or measuring probabilities is surprisingly difficult. Existing methods for conditional sampling at inference time seek mainly to enforce the constraints, which is insufficient to match the statistics of the distribution or compute the probability of the chosen events. To achieve these ends, optimally one would use the conditional score function, but its computation is typically intractable. In this work, we develop a probabilistic approximation scheme for the conditional score function which provably converges to the true distribution as the noise level decreases. With this scheme we are able to sample conditionally on nonlinear userdefined events at inference time, and matches data statistics even when sampling from the tails of the distribution.
UniCATS: A Unified Context-Aware Text-to-Speech Framework with Contextual VQ-Diffusion and Vocoding
Authors: Chenpeng Du, Yiwei Guo, Feiyu Shen, Zhijun Liu, Zheng Liang, Xie Chen, Shuai Wang, Hui Zhang, Kai Yu
Subjects: Sound (cs.SD); Audio and Speech Processing (eess.AS)
Abstract
The utilization of discrete speech tokens, divided into semantic tokens and acoustic tokens, has been proven superior to traditional acoustic feature mel-spectrograms in terms of naturalness and robustness for text-to-speech (TTS) synthesis. Recent popular models, such as VALL-E and SPEAR-TTS, allow zero-shot speaker adaptation through auto-regressive (AR) continuation of acoustic tokens extracted from a short speech prompt. However, these AR models are restricted to generate speech only in a left-to-right direction, making them unsuitable for speech editing where both preceding and following contexts are provided. Furthermore, these models rely on acoustic tokens, which have audio quality limitations imposed by the performance of audio codec models. In this study, we propose a unified context-aware TTS framework called UniCATS, which is capable of both speech continuation and editing. UniCATS comprises two components, an acoustic model CTX-txt2vec and a vocoder CTX-vec2wav. CTX-txt2vec employs contextual VQ-diffusion to predict semantic tokens from the input text, enabling it to incorporate the semantic context and maintain seamless concatenation with the surrounding context. Following that, CTX-vec2wav utilizes contextual vocoding to convert these semantic tokens into waveforms, taking into consideration the acoustic context. Our experimental results demonstrate that CTX-vec2wav outperforms HifiGAN and AudioLM in terms of speech resynthesis from semantic tokens. Moreover, we show that UniCATS achieves state-of-the-art performance in both speech continuation and editing.
Paste, Inpaint and Harmonize via Denoising: Subject-Driven Image Editing with Pre-Trained Diffusion Model
Abstract
Text-to-image generative models have attracted rising attention for flexible image editing via user-specified descriptions. However, text descriptions alone are not enough to elaborate the details of subjects, often compromising the subjects' identity or requiring additional per-subject fine-tuning. We introduce a new framework called \textit{Paste, Inpaint and Harmonize via Denoising} (PhD), which leverages an exemplar image in addition to text descriptions to specify user intentions. In the pasting step, an off-the-shelf segmentation model is employed to identify a user-specified subject within an exemplar image which is subsequently inserted into a background image to serve as an initialization capturing both scene context and subject identity in one. To guarantee the visual coherence of the generated or edited image, we introduce an inpainting and harmonizing module to guide the pre-trained diffusion model to seamlessly blend the inserted subject into the scene naturally. As we keep the pre-trained diffusion model frozen, we preserve its strong image synthesis ability and text-driven ability, thus achieving high-quality results and flexible editing with diverse texts. In our experiments, we apply PhD to both subject-driven image editing tasks and explore text-driven scene generation given a reference subject. Both quantitative and qualitative comparisons with baseline methods demonstrate that our approach achieves state-of-the-art performance in both tasks. More qualitative results can be found at \url{https://sites.google.com/view/phd-demo-page}.
Hyperbolic Graph Diffusion Model for Molecule Generation
Abstract
Recently, diffusion models have achieved remarkable performance in data generation, e.g., generating high-quality images. Nevertheless, chemistry molecules often have complex non-Euclidean spatial structures, with the behavior changing dynamically and unpredictably. Most existing diffusion models highly rely on computing the probability distribution, i.e., Gaussian distribution, in Euclidean space, which cannot capture internal non-Euclidean structures of molecules, especially the hierarchical structures of the implicit manifold surface represented by molecules. It has been observed that the complex hierarchical structures in hyperbolic embedding space become more prominent and easier to be captured. In order to leverage both the data generation power of diffusion models and the strong capability to extract complex geometric features of hyperbolic embedding, we propose to extend the diffusion model to hyperbolic manifolds for molecule generation, namely, Hyperbolic Graph Diffusion Model (HGDM). The proposed HGDM employs a hyperbolic variational autoencoder to generate the hyperbolic hidden representation of nodes and then a score-based hyperbolic graph neural network is used to learn the distribution in hyperbolic space. Numerical experimental results show that the proposed HGDM achieves higher performance on several molecular datasets, compared with state-of-the-art methods.
Viewset Diffusion: (0-)Image-Conditioned 3D Generative Models from 2D Data
Authors: Stanislaw Szymanowicz, Christian Rupprecht, Andrea Vedaldi
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
We present Viewset Diffusion: a framework for training image-conditioned 3D generative models from 2D data. Image-conditioned 3D generative models allow us to address the inherent ambiguity in single-view 3D reconstruction. Given one image of an object, there is often more than one possible 3D volume that matches the input image, because a single image never captures all sides of an object. Deterministic models are inherently limited to producing one possible reconstruction and therefore make mistakes in ambiguous settings. Modelling distributions of 3D shapes is challenging because 3D ground truth data is often not available. We propose to solve the issue of data availability by training a diffusion model which jointly denoises a multi-view image set.We constrain the output of Viewset Diffusion models to a single 3D volume per image set, guaranteeing consistent geometry. Training is done through reconstruction losses on renderings, allowing training with only three images per object. Our design of architecture and training scheme allows our model to perform 3D generation and generative, ambiguity-aware single-view reconstruction in a feed-forward manner. Project page: szymanowiczs.github.io/viewset-diffusion.
Rerender A Video: Zero-Shot Text-Guided Video-to-Video Translation
Abstract
Large text-to-image diffusion models have exhibited impressive proficiency in generating high-quality images. However, when applying these models to video domain, ensuring temporal consistency across video frames remains a formidable challenge. This paper proposes a novel zero-shot text-guided video-to-video translation framework to adapt image models to videos. The framework includes two parts: key frame translation and full video translation. The first part uses an adapted diffusion model to generate key frames, with hierarchical cross-frame constraints applied to enforce coherence in shapes, textures and colors. The second part propagates the key frames to other frames with temporal-aware patch matching and frame blending. Our framework achieves global style and local texture temporal consistency at a low cost (without re-training or optimization). The adaptation is compatible with existing image diffusion techniques, allowing our framework to take advantage of them, such as customizing a specific subject with LoRA, and introducing extra spatial guidance with ControlNet. Extensive experimental results demonstrate the effectiveness of our proposed framework over existing methods in rendering high-quality and temporally-coherent videos.
Keyword: adaptive
Multi-Platform Budget Management in Ad Markets with Non-IC Auctions
Abstract
In online advertising markets, budget-constrained advertisers acquire ad placements through repeated bidding in auctions on various platforms. We present a strategy for bidding optimally in a set of auctions that may or may not be incentive-compatible under the presence of budget constraints. Our strategy maximizes the expected total utility across auctions while satisfying the advertiser's budget constraints in expectation. Additionally, we investigate the online setting where the advertiser must submit bids across platforms while learning about other bidders' bids over time. Our algorithm has $O(T^{3/4})$ regret under the full-information setting. Finally, we demonstrate that our algorithms have superior cumulative regret on both synthetic and real-world datasets of ad placement auctions, compared to existing adaptive pacing algorithms.
Reviving Shift Equivariance in Vision Transformers
Authors: Peijian Ding, Davit Soselia, Thomas Armstrong, Jiahao Su, Furong Huang
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Abstract
Shift equivariance is a fundamental principle that governs how we perceive the world - our recognition of an object remains invariant with respect to shifts. Transformers have gained immense popularity due to their effectiveness in both language and vision tasks. While the self-attention operator in vision transformers (ViT) is permutation-equivariant and thus shift-equivariant, patch embedding, positional encoding, and subsampled attention in ViT variants can disrupt this property, resulting in inconsistent predictions even under small shift perturbations. Although there is a growing trend in incorporating the inductive bias of convolutional neural networks (CNNs) into vision transformers, it does not fully address the issue. We propose an adaptive polyphase anchoring algorithm that can be seamlessly integrated into vision transformer models to ensure shift-equivariance in patch embedding and subsampled attention modules, such as window attention and global subsampled attention. Furthermore, we utilize depth-wise convolution to encode positional information. Our algorithms enable ViT, and its variants such as Twins to achieve 100% consistency with respect to input shift, demonstrate robustness to cropping, flipping, and affine transformations, and maintain consistent predictions even when the original models lose 20 percentage points on average when shifted by just a few pixels with Twins' accuracy dropping from 80.57% to 62.40%.
Heterophily-aware Social Bot Detection with Supervised Contrastive Learning
Abstract
Detecting ever-evolving social bots has become increasingly challenging. Advanced bots tend to interact more with humans as a camouflage to evade detection. While graph-based detection methods can exploit various relations in social networks to model node behaviors, the aggregated information from neighbors largely ignore the inherent heterophily, i.e., the connections between different classes of accounts. Message passing mechanism on heterophilic edges can lead to feature mixture between bots and normal users, resulting in more false negatives. In this paper, we present BotSCL, a heterophily-aware contrastive learning framework that can adaptively differentiate neighbor representations of heterophilic relations while assimilating the representations of homophilic neighbors. Specifically, we employ two graph augmentation methods to generate different graph views and design a channel-wise and attention-free encoder to overcome the limitation of neighbor information summing. Supervised contrastive learning is used to guide the encoder to aggregate class-specific information. Extensive experiments on two social bot detection benchmarks demonstrate that BotSCL outperforms baseline approaches including the state-of-the-art bot detection approaches, partially heterophilic GNNs and self-supervised contrast learning methods.
Retrieve Anyone: A General-purpose Person Re-identification Task with Instructions
Abstract
Human intelligence can retrieve any person according to both visual and language descriptions. However, the current computer vision community studies specific person re-identification (ReID) tasks in different scenarios separately, which limits the applications in the real world. This paper strives to resolve this problem by proposing a new instruct-ReID task that requires the model to retrieve images according to the given image or language instructions.Our instruct-ReID is a more general ReID setting, where existing ReID tasks can be viewed as special cases by designing different instructions. We propose a large-scale OmniReID benchmark and an adaptive triplet loss as a baseline method to facilitate research in this new setting. Experimental results show that the baseline model trained on our OmniReID benchmark can improve +0.5%, +3.3% mAP on Market1501 and CUHK03 for traditional ReID, +2.1%, +0.2%, +15.3% mAP on PRCC, VC-Clothes, LTCC for clothes-changing ReID, +12.5% mAP on COCAS+ real2 for clothestemplate based clothes-changing ReID when using only RGB images, +25.5% mAP on COCAS+ real2 for our newly defined language-instructed ReID. The dataset, model, and code will be available at https://github.com/hwz-zju/Instruct-ReID.
A Simple Unified Uncertainty-Guided Framework for Offline-to-Online Reinforcement Learning
Abstract
Offline reinforcement learning (RL) provides a promising solution to learning an agent fully relying on a data-driven paradigm. However, constrained by the limited quality of the offline dataset, its performance is often sub-optimal. Therefore, it is desired to further finetune the agent via extra online interactions before deployment. Unfortunately, offline-to-online RL can be challenging due to two main challenges: constrained exploratory behavior and state-action distribution shift. To this end, we propose a Simple Unified uNcertainty-Guided (SUNG) framework, which naturally unifies the solution to both challenges with the tool of uncertainty. Specifically, SUNG quantifies uncertainty via a VAE-based state-action visitation density estimator. To facilitate efficient exploration, SUNG presents a practical optimistic exploration strategy to select informative actions with both high value and high uncertainty. Moreover, SUNG develops an adaptive exploitation method by applying conservative offline RL objectives to high-uncertainty samples and standard online RL objectives to low-uncertainty samples to smoothly bridge offline and online stages. SUNG achieves state-of-the-art online finetuning performance when combined with different offline RL methods, across various environments and datasets in D4RL benchmark.
Fixed-Budget Best-Arm Identification with Heterogeneous Reward Variances
Abstract
We study the problem of best-arm identification (BAI) in the fixed-budget setting with heterogeneous reward variances. We propose two variance-adaptive BAI algorithms for this setting: SHVar for known reward variances and SHAdaVar for unknown reward variances. Our algorithms rely on non-uniform budget allocations among the arms where the arms with higher reward variances are pulled more often than those with lower variances. The main algorithmic novelty is in the design of SHAdaVar, which allocates budget greedily based on overestimating the unknown reward variances. We bound probabilities of misidentifying the best arms in both SHVar and SHAdaVar. Our analyses rely on novel lower bounds on the number of pulls of an arm that do not require closed-form solutions to the budget allocation problem. Since one of our budget allocation problems is analogous to the optimal experiment design with unknown variances, we believe that our results are of a broad interest. Our experiments validate our theory, and show that SHVar and SHAdaVar outperform algorithms from prior works with analytical guarantees.
UOD: Universal One-shot Detection of Anatomical Landmarks
Abstract
One-shot medical landmark detection gains much attention and achieves great success for its label-efficient training process. However, existing one-shot learning methods are highly specialized in a single domain and suffer domain preference heavily in the situation of multi-domain unlabeled data. Moreover, one-shot learning is not robust that it faces performance drop when annotating a sub-optimal image. To tackle these issues, we resort to developing a domain-adaptive one-shot landmark detection framework for handling multi-domain medical images, named Universal One-shot Detection (UOD). UOD consists of two stages and two corresponding universal models which are designed as combinations of domain-specific modules and domain-shared modules. In the first stage, a domain-adaptive convolution model is self-supervised learned to generate pseudo landmark labels. In the second stage, we design a domain-adaptive transformer to eliminate domain preference and build the global context for multi-domain data. Even though only one annotated sample from each domain is available for training, the domain-shared modules help UOD aggregate all one-shot samples to detect more robust and accurate landmarks. We investigated both qualitatively and quantitatively the proposed UOD on three widely-used public X-ray datasets in different anatomical domains (i.e., head, hand, chest) and obtained state-of-the-art performances in each domain.
Dynamically Masked Discriminator for Generative Adversarial Networks
Abstract
Training Generative Adversarial Networks (GANs) remains a challenging problem. The discriminator trains the generator by learning the distribution of real/generated data. However, the distribution of generated data changes throughout the training process, which is difficult for the discriminator to learn. In this paper, we propose a novel method for GANs from the viewpoint of online continual learning. We observe that the discriminator model, trained on historically generated data, often slows down its adaptation to the changes in the new arrival generated data, which accordingly decreases the quality of generated results. By treating the generated data in training as a stream, we propose to detect whether the discriminator slows down the learning of new knowledge in generated data. Therefore, we can explicitly enforce the discriminator to learn new knowledge fast. Particularly, we propose a new discriminator, which automatically detects its retardation and then dynamically masks its features, such that the discriminator can adaptively learn the temporally-vary distribution of generated data. Experimental results show our method outperforms the state-of-the-art approaches.
Contextual Dictionary Lookup for Knowledge Graph Completion
Abstract
Knowledge graph completion (KGC) aims to solve the incompleteness of knowledge graphs (KGs) by predicting missing links from known triples, numbers of knowledge graph embedding (KGE) models have been proposed to perform KGC by learning embeddings. Nevertheless, most existing embedding models map each relation into a unique vector, overlooking the specific fine-grained semantics of them under different entities. Additionally, the few available fine-grained semantic models rely on clustering algorithms, resulting in limited performance and applicability due to the cumbersome two-stage training process. In this paper, we present a novel method utilizing contextual dictionary lookup, enabling conventional embedding models to learn fine-grained semantics of relations in an end-to-end manner. More specifically, we represent each relation using a dictionary that contains multiple latent semantics. The composition of a given entity and the dictionary's central semantics serves as the context for generating a lookup, thus determining the fine-grained semantics of the relation adaptively. The proposed loss function optimizes both the central and fine-grained semantics simultaneously to ensure their semantic consistency. Besides, we introduce two metrics to assess the validity and accuracy of the dictionary lookup operation. We extend several KGE models with the method, resulting in substantial performance improvements on widely-used benchmark datasets.
Stepsize Learning for Policy Gradient Methods in Contextual Markov Decision Processes
Authors: Luca Sabbioni, Francesco Corda, Marcello Restelli
Abstract
Policy-based algorithms are among the most widely adopted techniques in model-free RL, thanks to their strong theoretical groundings and good properties in continuous action spaces. Unfortunately, these methods require precise and problem-specific hyperparameter tuning to achieve good performance, and tend to struggle when asked to accomplish a series of heterogeneous tasks. In particular, the selection of the step size has a crucial impact on their ability to learn a highly performing policy, affecting the speed and the stability of the training process, and often being the main culprit for poor results. In this paper, we tackle these issues with a Meta Reinforcement Learning approach, by introducing a new formulation, known as meta-MDP, that can be used to solve any hyperparameter selection problem in RL with contextual processes. After providing a theoretical Lipschitz bound to the difference of performance in different tasks, we adopt the proposed framework to train a batch RL algorithm to dynamically recommend the most adequate step size for different policies and tasks. In conclusion, we present an experimental campaign to show the advantages of selecting an adaptive learning rate in heterogeneous environments.
Inferring dynamic regulatory interaction graphs from time series data with perturbations
Authors: Dhananjay Bhaskar, Sumner Magruder, Edward De Brouwer, Aarthi Venkat, Frederik Wenkel, Guy Wolf, Smita Krishnaswamy
Abstract
Complex systems are characterized by intricate interactions between entities that evolve dynamically over time. Accurate inference of these dynamic relationships is crucial for understanding and predicting system behavior. In this paper, we propose Regulatory Temporal Interaction Network Inference (RiTINI) for inferring time-varying interaction graphs in complex systems using a novel combination of space-and-time graph attentions and graph neural ordinary differential equations (ODEs). RiTINI leverages time-lapse signals on a graph prior, as well as perturbations of signals at various nodes in order to effectively capture the dynamics of the underlying system. This approach is distinct from traditional causal inference networks, which are limited to inferring acyclic and static graphs. In contrast, RiTINI can infer cyclic, directed, and time-varying graphs, providing a more comprehensive and accurate representation of complex systems. The graph attention mechanism in RiTINI allows the model to adaptively focus on the most relevant interactions in time and space, while the graph neural ODEs enable continuous-time modeling of the system's dynamics. We evaluate RiTINI's performance on various simulated and real-world datasets, demonstrating its state-of-the-art capability in inferring interaction graphs compared to previous methods.
PSSTRNet: Progressive Segmentation-guided Scene Text Removal Network
Authors: Guangtao Lyu, Anna Zhu
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Scene text removal (STR) is a challenging task due to the complex text fonts, colors, sizes, and background textures in scene images. However, most previous methods learn both text location and background inpainting implicitly within a single network, which weakens the text localization mechanism and makes a lossy background. To tackle these problems, we propose a simple Progressive Segmentation-guided Scene Text Removal Network(PSSTRNet) to remove the text in the image iteratively. It contains two decoder branches, a text segmentation branch, and a text removal branch, with a shared encoder. The text segmentation branch generates text mask maps as the guidance for the regional removal branch. In each iteration, the original image, previous text removal result, and text mask are input to the network to extract the rest part of the text segments and cleaner text removal result. To get a more accurate text mask map, an update module is developed to merge the mask map in the current and previous stages. The final text removal result is obtained by adaptive fusion of results from all previous stages. A sufficient number of experiments and ablation studies conducted on the real and synthetic public datasets demonstrate our proposed method achieves state-of-the-art performance. The source code of our work is available at: \href{https://github.com/GuangtaoLyu/PSSTRNet}{https://github.com/GuangtaoLyu/PSSTRNet.}
Show me the numbers! -- Student-facing Interventions in Adaptive Learning Environments for German Spelling
Abstract
Since adaptive learning comes in many shapes and sizes, it is crucial to find out which adaptations can be meaningful for which areas of learning. Our work presents the result of an experiment conducted on an online platform for the acquisition of German spelling skills. We compared the traditional online learning platform to three different adaptive versions of the platform that implement machine learning-based student-facing interventions that show the personalized solution probability. We evaluate the different interventions with regard to the error rate, the number of early dropouts, and the users competency. Our results show that the number of mistakes decreased in comparison to the control group. Additionally, an increasing number of dropouts was found. We did not find any significant effects on the users competency. We conclude that student-facing adaptive learning environments are effective in improving a persons error rate and should be chosen wisely to have a motivating impact.
Keyword: quantization
GQFedWAvg: Optimization-Based Quantized Federated Learning in General Edge Computing Systems
Authors: Yangchen Li, Ying Cui, Vincent Lau
Subjects: Machine Learning (cs.LG); Distributed, Parallel, and Cluster Computing (cs.DC)
Abstract
The optimal implementation of federated learning (FL) in practical edge computing systems has been an outstanding problem. In this paper, we propose an optimization-based quantized FL algorithm, which can appropriately fit a general edge computing system with uniform or nonuniform computing and communication resources at the workers. Specifically, we first present a new random quantization scheme and analyze its properties. Then, we propose a general quantized FL algorithm, namely GQFedWAvg. Specifically, GQFedWAvg applies the proposed quantization scheme to quantize wisely chosen model update-related vectors and adopts a generalized mini-batch stochastic gradient descent (SGD) method with the weighted average local model updates in global model aggregation. Besides, GQFedWAvg has several adjustable algorithm parameters to flexibly adapt to the computing and communication resources at the server and workers. We also analyze the convergence of GQFedWAvg. Next, we optimize the algorithm parameters of GQFedWAvg to minimize the convergence error under the time and energy constraints. We successfully tackle the challenging non-convex problem using general inner approximation (GIA) and multiple delicate tricks. Finally, we interpret GQFedWAvg's function principle and show its considerable gains over existing FL algorithms using numerical results.
SqueezeLLM: Dense-and-Sparse Quantization
Authors: Sehoon Kim, Coleman Hooper, Amir Gholami, Zhen Dong, Xiuyu Li, Sheng Shen, Michael W. Mahoney, Kurt Keutzer
Subjects: Computation and Language (cs.CL); Machine Learning (cs.LG)
Abstract
Generative Large Language Models (LLMs) have demonstrated remarkable results for a wide range of tasks. However, deploying these models for inference has been a significant challenge due to their unprecedented resource requirements. This has forced existing deployment frameworks to use multi-GPU inference pipelines, which are often complex and costly, or to use smaller and less performant models. In this work, we demonstrate that the main bottleneck for generative inference with LLMs is memory bandwidth, rather than compute, specifically for single batch inference. While quantization has emerged as a promising solution by representing model weights with reduced precision, previous efforts have often resulted in notable performance degradation. To address this, we introduce SqueezeLLM, a post-training quantization framework that not only enables lossless compression to ultra-low precisions of up to 3-bit, but also achieves higher quantization performance under the same memory constraint. Our framework incorporates two novel ideas: (i) sensitivity-based non-uniform quantization, which searches for the optimal bit precision assignment based on second-order information; and (ii) the Dense-and-Sparse decomposition that stores outliers and sensitive weight values in an efficient sparse format. When applied to the LLaMA models, our 3-bit quantization significantly reduces the perplexity gap from the FP16 baseline by up to 2.1x as compared to the state-of-the-art methods with the same memory requirement. Furthermore, when deployed on an A6000 GPU, our quantized models achieve up to 2.3x speedup compared to the baseline. Our code is open-sourced and available online.
Abstract
In this work, we addresses the problem of modeling distributions of graphs. We introduce the Vector-Quantized Graph Auto-Encoder (VQ-GAE), a permutation-equivariant discrete auto-encoder and designed to model the distribution of graphs. By exploiting the permutation-equivariance of graph neural networks (GNNs), our autoencoder circumvents the problem of the ordering of the graph representation. We leverage the capability of GNNs to capture local structures of graphs while employing vector-quantization to prevent the mapping of discrete objects to a continuous latent space. Furthermore, the use of autoregressive models enables us to capture the global structure of graphs via the latent representation. We evaluate our model on standard datasets used for graph generation and observe that it achieves excellent performance on some of the most salient evaluation metrics compared to the state-of-the-art.
Keyword: efficient
Value function estimation using conditional diffusion models for control
Expressivity Enhancement with Efficient Quadratic Neurons for Convolutional Neural Networks
Referring to Screen Texts with Voice Assistants
Novel Regression and Least Square Support Vector Machine Learning Technique for Air Pollution Forecasting
A Holistic Approach to Unifying Automatic Concept Extraction and Concept Importance Estimation
Online Prototype Alignment for Few-shot Policy Transfer
A New Probabilistic Distance Metric With Application In Gaussian Mixture Reduction
Exploratory analysis of a measurement scale of an information security management system
Set-based state estimation and fault diagnosis using constrained zonotopes and applications
Robust Reinforcement Learning through Efficient Adversarial Herding
Instant Multi-View Head Capture through Learnable Registration
Constructing Printable Surfaces with View-Dependent Appearance
ELF Codes: Concatenated Codes with an Expurgating Linear Function as the Outer Code
Multi-objective Molecular Optimization for Opioid Use Disorder Treatment Using Generative Network Complex
Improving Opinion-based Question Answering Systems Through Label Error Detection and Overwrite
PaVa: a novel Path-based Valley-seeking clustering algorithm
Energy Efficient RAN Slicing and Beams Selection for Multiplexing of Heterogeneous Services in 5G mmWave Networks
Using Collision Momentum in Deep Reinforcement Learning Based Adversarial Pedestrian Modeling
Hybrid and Oriented Harmonic Potentials for Safe Task Execution in Unknown Environment
A Simple Unified Uncertainty-Guided Framework for Offline-to-Online Reinforcement Learning
HAUSER: Towards Holistic and Automatic Evaluation of Simile Generation
Ethical Aspects of ChatGPT in Software Engineering Research
Action Recognition with Multi-stream Motion Modeling and Mutual Information Maximization
Parametric Implicit Face Representation for Audio-Driven Facial Reenactment
Binary Radiance Fields
Invertible Bloom Lookup Tables with Less Memory and Less Randomness
Practice with Graph-based ANN Algorithms on Sparse Data: Chi-square Two-tower model, HNSW, Sign Cauchy Projections
handcrafted'' features are often sparse. Embedding vectors from trained models can also be very sparse, for example, embeddings trained via the
ReLu'' activation function. In this paper, we report our exploration of efficient search in sparse data with graph-based ANN algorithms (e.g., HNSW, or SONG which is the GPU version of HNSW), which are popular in industrial practice, e.g., search and ads (advertising). We experiment with the proprietary ads targeting application, as well as benchmark public datasets. For ads targeting, we train embeddings with the standardcosine two-tower'' model and we also develop the
chi-square two-tower'' model. Both models produce (highly) sparse embeddings when they are integrated with theReLu'' activation function. In EBR (embedding-based retrieval) applications, after we the embeddings are trained, the next crucial task is the approximate near neighbor (ANN) search for serving. While there are many ANN algorithms we can choose from, in this study, we focus on the graph-based ANN algorithm (e.g., HNSW-type). Sparse embeddings should help improve the efficiency of EBR. One benefit is the reduced memory cost for the embeddings. The other obvious benefit is the reduced computational time for evaluating similarities, because, for graph-based ANN algorithms such as HNSW, computing similarities is often the dominating cost. In addition to the effort on leveraging data sparsity for storage and computation, we also integrate
sign cauchy random projections'' (SignCRP) to hash vectors to bits, to further reduce the memory cost and speed up the ANN search. In NIPS'13, SignCRP was proposed to hash the chi-square similarity, which is a well-adopted nonlinear kernel in NLP and computer vision. Therefore, the chi-square two-tower model, SignCRP, and HNSW are now tightly integrated.UOD: Universal One-shot Detection of Anatomical Landmarks
SqueezeLLM: Dense-and-Sparse Quantization
Sea Ice Segmentation From SAR Data by Convolutional Transformer Networks
Automating Microservices Test Failure Analysis using Kubernetes Cluster Logs
Few-shot Multi-domain Knowledge Rearming for Context-aware Defence against Advanced Persistent Threats
E2E-LOAD: End-to-End Long-form Online Action Detection
Just a Second -- Scheduling Thousands of Time-Triggered Streams in Large-Scale Networks
Kernelized Reinforcement Learning with Order Optimal Regret Bounds
NAVER LABS Europe's Multilingual Speech Translation Systems for the IWSLT 2023 Low-Resource Track
A Cloud-based Machine Learning Pipeline for the Efficient Extraction of Insights from Customer Reviews
Efficient GPU implementation of a class of array permutations
GEmo-CLAP: Gender-Attribute-Enhanced Contrastive Language-Audio Pretraining for Speech Emotion Recognition
Expanding the Scope of DAWN: A Novel Version for Weighted Shortest Path Problem
Robustly Learning a Single Neuron via Sharpness
A dual number formulation to efficiently compute higher order directional derivatives
WebGLM: Towards An Efficient Web-Enhanced Question Answering System with Human Preferences
Continuous Cost Aggregation for Dual-Pixel Disparity Extraction
Oracle-Efficient Pessimism: Offline Policy Optimization in Contextual Bandits
GPT-Calls: Enhancing Call Segmentation and Tagging by Generating Synthetic Conversations via Large Language Models
Parting with Misconceptions about Learning-based Vehicle Motion Planning
One-for-All: Generalized LoRA for Parameter-Efficient Fine-tuning
Keyword: faster
A New Probabilistic Distance Metric With Application In Gaussian Mixture Reduction
FIRE: An Optimization Approach for Fast Interpretable Rule Extraction
3D molecule generation by denoising voxel grids
Learning Unnormalized Statistical Models via Compositional Optimization
Galactic: Scaling End-to-End Reinforcement Learning for Rearrangement at 100k Steps-Per-Second
Few-shot Multi-domain Knowledge Rearming for Context-aware Defence against Advanced Persistent Threats
E2E-LOAD: End-to-End Long-form Online Action Detection
DreamDecompiler: Improved Bayesian Program Learning by Decompiling Amortised Knowledge
CAMEO: A Causal Transfer Learning Approach for Performance Optimization of Configurable Computer Systems
Keyword: mobile
Ethical Considerations of AR Applications in Smartphones; A Systematic Literature Review of Consumer Perspectives
Learning Any-View 6DoF Robotic Grasping in Cluttered Scenes via Neural Surface Rendering
SoK: Decoding the Super App Enigma: The Security Mechanisms, Threats, and Trade-offs in OS-alike Apps
Energy Efficient RAN Slicing and Beams Selection for Multiplexing of Heterogeneous Services in 5G mmWave Networks
Galactic: Scaling End-to-End Reinforcement Learning for Rearrangement at 100k Steps-Per-Second
Keyword: pruning
There is no result
Keyword: diffusion
Value function estimation using conditional diffusion models for control
G-invariant diffusion maps
3D molecule generation by denoising voxel grids
Multi-objective Molecular Optimization for Opioid Use Disorder Treatment Using Generative Network Complex
User-defined Event Sampling and Uncertainty Quantification in Diffusion Models for Physical Dynamical Systems
UniCATS: A Unified Context-Aware Text-to-Speech Framework with Contextual VQ-Diffusion and Vocoding
Paste, Inpaint and Harmonize via Denoising: Subject-Driven Image Editing with Pre-Trained Diffusion Model
Hyperbolic Graph Diffusion Model for Molecule Generation
Viewset Diffusion: (0-)Image-Conditioned 3D Generative Models from 2D Data
Rerender A Video: Zero-Shot Text-Guided Video-to-Video Translation
Keyword: adaptive
Multi-Platform Budget Management in Ad Markets with Non-IC Auctions
Reviving Shift Equivariance in Vision Transformers
Heterophily-aware Social Bot Detection with Supervised Contrastive Learning
Retrieve Anyone: A General-purpose Person Re-identification Task with Instructions
A Simple Unified Uncertainty-Guided Framework for Offline-to-Online Reinforcement Learning
Fixed-Budget Best-Arm Identification with Heterogeneous Reward Variances
UOD: Universal One-shot Detection of Anatomical Landmarks
Dynamically Masked Discriminator for Generative Adversarial Networks
Contextual Dictionary Lookup for Knowledge Graph Completion
Stepsize Learning for Policy Gradient Methods in Contextual Markov Decision Processes
Inferring dynamic regulatory interaction graphs from time series data with perturbations
PSSTRNet: Progressive Segmentation-guided Scene Text Removal Network
Show me the numbers! -- Student-facing Interventions in Adaptive Learning Environments for German Spelling
Keyword: quantization
GQFedWAvg: Optimization-Based Quantized Federated Learning in General Edge Computing Systems
SqueezeLLM: Dense-and-Sparse Quantization
Vector-Quantized Graph Auto-Encoder