Abstract
While alignment algorithms are now commonly used to tune pre-trained language models towards a user's preferences, we lack explanations for the underlying mechanisms in which models become ``aligned'', thus making it difficult to explain phenomena like jailbreaks. In this work we study a popular algorithm, direct preference optimization (DPO), and the mechanisms by which it reduces toxicity. Namely, we first study how toxicity is represented and elicited in a pre-trained language model, GPT2-medium. We then apply DPO with a carefully crafted pairwise dataset to reduce toxicity. We examine how the resulting model averts toxic outputs, and find that capabilities learned from pre-training are not removed, but rather bypassed. We use this insight to demonstrate a simple method to un-align the model, reverting it back to its toxic behavior.
Beyond Regrets: Geometric Metrics for Bayesian Optimization
Abstract
Bayesian optimization is a principled optimization strategy for a black-box objective function. It shows its effectiveness in a wide variety of real-world applications such as scientific discovery and experimental design. In general, the performance of Bayesian optimization is assessed by regret-based metrics such as instantaneous, simple, and cumulative regrets. These metrics only rely on function evaluations, so that they do not consider geometric relationships between query points and global solutions, or query points themselves. Notably, they cannot discriminate if multiple global solutions are successfully found. Moreover, they do not evaluate Bayesian optimization's abilities to exploit and explore a search space given. To tackle these issues, we propose four new geometric metrics, i.e., precision, recall, average degree, and average distance. These metrics allow us to compare Bayesian optimization algorithms considering the geometry of both query points and global optima, or query points. However, they are accompanied by an extra parameter, which needs to be carefully determined. We therefore devise the parameter-free forms of the respective metrics by integrating out the additional parameter. Finally, we empirically validate that our proposed metrics can provide more convincing interpretation and understanding of Bayesian optimization algorithms from distinct perspectives, compared to the conventional metrics.
Decentralized Multi-Task Online Convex Optimization Under Random Link Failures
Authors: Authors: Wenjing Yan, Xuanyu Cao
Subjects: Machine Learning (cs.LG); Optimization and Control (math.OC)
Abstract
Decentralized optimization methods often entail information exchange between neighbors. Transmission failures can happen due to network congestion, hardware/software issues, communication outage, and other factors. In this paper, we investigate the random link failure problem in decentralized multi-task online convex optimization, where agents have individual decisions that are coupled with each other via pairwise constraints. Although widely used in constrained optimization, conventional saddle-point algorithms are not directly applicable here because of random packet dropping. To address this issue, we develop a robust decentralized saddle-point algorithm against random link failures with heterogeneous probabilities by replacing the missing decisions of neighbors with their latest received values. Then, by judiciously bounding the accumulated deviation stemming from this replacement, we first establish that our algorithm achieves $\mathcal{O}(\sqrt{T})$ regret and $\mathcal{O}(T^\frac{3}{4})$ constraint violations for the full information scenario, where the complete information on the local cost function is revealed to each agent at the end of each time slot. These two bounds match, in order sense, the performance bounds of algorithms with perfect communications. Further, we extend our algorithm and analysis to the two-point bandit feedback scenario, where only the values of the local cost function at two random points are disclosed to each agent sequentially. Performance bounds of the same orders as the full information case are derived. Finally, we corroborate the efficacy of the proposed algorithms and the analytical results through numerical simulations.
Fast & Fair: Efficient Second-Order Robust Optimization for Fairness in Machine Learning
Authors: Authors: Allen Minch, Hung Anh Vu, Anne Marie Warren
Subjects: Machine Learning (cs.LG); Computers and Society (cs.CY); Numerical Analysis (math.NA)
Abstract
This project explores adversarial training techniques to develop fairer Deep Neural Networks (DNNs) to mitigate the inherent bias they are known to exhibit. DNNs are susceptible to inheriting bias with respect to sensitive attributes such as race and gender, which can lead to life-altering outcomes (e.g., demographic bias in facial recognition software used to arrest a suspect). We propose a robust optimization problem, which we demonstrate can improve fairness in several datasets, both synthetic and real-world, using an affine linear model. Leveraging second order information, we are able to find a solution to our optimization problem more efficiently than a purely first order method.
From Function to Distribution Modeling: A PAC-Generative Approach to Offline Optimization
Authors: Authors: Qiang Zhang, Ruida Zhou, Yang Shen, Tie Liu
Abstract
This paper considers the problem of offline optimization, where the objective function is unknown except for a collection of offline" data examples. While recent years have seen a flurry of work on applying various machine learning techniques to the offline optimization problem, the majority of these work focused on learning a surrogate of the unknown objective function and then applying existing optimization algorithms. While the idea of modeling the unknown objective function is intuitive and appealing, from the learning point of view it also makes it very difficult to tune the objective of the learner according to the objective of optimization. Instead of learning and then optimizing the unknown objective function, in this paper we take on a less intuitive but more direct view that optimization can be thought of as a process of sampling from a generative model. To learn an effective generative model from the offline data examples, we consider the standard technique ofre-weighting", and our main technical contribution is a probably approximately correct (PAC) lower bound on the natural optimization objective, which allows us to jointly learn a weight function and a score-based generative model. The robustly competitive performance of the proposed approach is demonstrated via empirical studies using the standard offline optimization benchmarks.
Understanding LLMs: A Comprehensive Overview from Training to Inference
Authors: Authors: Yiheng Liu, Hao He, Tianle Han, Xu Zhang, Mengyuan Liu, Jiaming Tian, Yutong Zhang, Jiaqi Wang, Xiaohui Gao, Tianyang Zhong, Yi Pan, Shaochen Xu, Zihao Wu, Zhengliang Liu, Xin Zhang, Shu Zhang, Xintao Hu, Tuo Zhang, Ning Qiang, Tianming Liu, Bao Ge
Abstract
The introduction of ChatGPT has led to a significant increase in the utilization of Large Language Models (LLMs) for addressing downstream tasks. There's an increasing focus on cost-efficient training and deployment within this context. Low-cost training and deployment of LLMs represent the future development trend. This paper reviews the evolution of large language model training techniques and inference deployment technologies aligned with this emerging trend. The discussion on training includes various aspects, including data preprocessing, training architecture, pre-training tasks, parallel training, and relevant content related to model fine-tuning. On the inference side, the paper covers topics such as model compression, parallel computation, memory scheduling, and structural optimization. It also explores LLMs' utilization and provides insights into their future development.
Joint Beamforming and Offloading Design for Integrated Sensing, Communication and Computation System
Authors: Authors: Peng Liu, Zesong Fei, Xinyi Wang, Yiqing Zhou, Yan Zhang, Fan Liu
Subjects: Information Theory (cs.IT); Signal Processing (eess.SP)
Abstract
Mobile edge computing (MEC) is powerful to alleviate the heavy computing tasks in integrated sensing and communication (ISAC) systems. In this paper, we investigate joint beamforming and offloading design in a three-tier integrated sensing, communication and computation (ISCC) framework comprising one cloud server, multiple mobile edge servers, and multiple terminals. While executing sensing tasks, the user terminals can optionally offload sensing data to either MEC server or cloud servers. To minimize the execution latency, we jointly optimize the transmit beamforming matrices and offloading decision variables under the constraint of sensing performance. An alternating optimization algorithm based on multidimensional fractional programming is proposed to tackle the non-convex problem. Simulation results demonstrates the superiority of the proposed mechanism in terms of convergence and task execution latency reduction, compared with the state-of-the-art two-tier ISCC framework.
ICE-GRT: Instruction Context Enhancement by Generative Reinforcement based Transformers
Authors: Authors: Chen Zheng, Ke Sun, Da Tang, Yukun Ma, Yuyu Zhang, Chenguang Xi, Xun Zhou
Abstract
The emergence of Large Language Models (LLMs) such as ChatGPT and LLaMA encounter limitations in domain-specific tasks, with these models often lacking depth and accuracy in specialized areas, and exhibiting a decrease in general capabilities when fine-tuned, particularly analysis ability in small sized models. To address these gaps, we introduce ICE-GRT, utilizing Reinforcement Learning from Human Feedback (RLHF) grounded in Proximal Policy Optimization (PPO), demonstrating remarkable ability in in-domain scenarios without compromising general task performance. Our exploration of ICE-GRT highlights its understanding and reasoning ability to not only generate robust answers but also to provide detailed analyses of the reasons behind the answer. This capability marks a significant progression beyond the scope of Supervised Fine-Tuning models. The success of ICE-GRT is dependent on several crucial factors, including Appropriate Data, Reward Size Scaling, KL-Control, Advantage Normalization, etc. The ICE-GRT model exhibits state-of-the-art performance in domain-specific tasks and across 12 general Language tasks against equivalent size and even larger size LLMs, highlighting the effectiveness of our approach. We provide a comprehensive analysis of the ICE-GRT, underscoring the significant advancements it brings to the field of LLM.
Performance Trade-off and Joint Waveform Design for MIMO-OFDM DFRC Systems
Authors: Authors: Tianchen Liu, Liang Wu, Bo An, Zaichen Zhang, Jian Dang, Jiangzhou Wang
Subjects: Information Theory (cs.IT); Signal Processing (eess.SP)
Abstract
Dual-functional radar-communication (DFRC) has attracted considerable attention. This paper considers the frequency-selective multipath fading environment and proposes DFRC waveform design strategies based on multiple-input and multiple-output (MIMO) and orthogonal frequency division multiplexing (OFDM) techniques. In the proposed waveform design strategies, the Cramer-Rao bound (CRB) of the radar system, the inter-stream interference (ISI) and the achievable rate of the communication system, are respectively considered as the performance metrics. In this paper, we focus on the performance trade-off between the radar system and the communication system, and the optimization problems are formulated. In the ISI minimization based waveform design strategy, the optimization problem is convex and can be easily solved. In the achievable rate maximization based waveform design strategy, we propose a water-filling (WF) and sequential quadratic programming (SQP) based algorithm to derive the covariance matrix and the precoding matrix. Simulation results validate the proposed DFRC waveform designs and show that the achievable rate maximization based strategy has a better performance than the ISI minimization based strategy.
Abstract
Generating explanations for graph neural networks (GNNs) has been studied to understand their behavior in analytical tasks such as graph classification. Existing approaches aim to understand the overall results of GNNs rather than providing explanations for specific class labels of interest, and may return explanation structures that are hard to access, nor directly queryable. We propose GVEX, a novel paradigm that generates Graph Views for EXplanation. (1) We design a two-tier explanation structure called explanation views. An explanation view consists of a set of graph patterns and a set of induced explanation subgraphs. Given a database G of multiple graphs and a specific class label l assigned by a GNN-based classifier M, it concisely describes the fraction of G that best explains why l is assigned by M. (2) We propose quality measures and formulate an optimization problem to compute optimal explanation views for GNN explanation. We show that the problem is $\Sigma^2_P$-hard. (3) We present two algorithms. The first one follows an explain-and-summarize strategy that first generates high-quality explanation subgraphs which best explain GNNs in terms of feature influence maximization, and then performs a summarization step to generate patterns. We show that this strategy provides an approximation ratio of 1/2. Our second algorithm performs a single-pass to an input node stream in batches to incrementally maintain explanation views, having an anytime quality guarantee of 1/4 approximation. Using real-world benchmark data, we experimentally demonstrate the effectiveness, efficiency, and scalability of GVEX. Through case studies, we showcase the practical applications of GVEX.
PEFT for Speech: Unveiling Optimal Placement, Merging Strategies, and Ensemble Techniques
Abstract
Parameter-Efficient Fine-Tuning (PEFT) is increasingly recognized as an effective method in speech processing. However, the optimal approach and the placement of PEFT methods remain inconclusive. Our study conducts extensive experiments to compare different PEFT methods and their layer-wise placement adapting Differentiable Architecture Search (DARTS). We also explore the use of ensemble learning to leverage diverse PEFT strategies. The results reveal that DARTS does not outperform the baseline approach, which involves inserting the same PEFT method into all layers of a Self-Supervised Learning (SSL) model. In contrast, an ensemble learning approach, particularly one employing majority voting, demonstrates superior performance. Our statistical evidence indicates that different PEFT methods learn in varied ways. This variation might explain why the synergistic integration of various PEFT methods through ensemble learning can harness their unique learning capabilities more effectively compared to individual layer-wise optimization.
Marginal Debiased Network for Fair Visual Recognition
Authors: Authors: Mei Wang, Weihong Deng, Sen Su
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Deep neural networks (DNNs) are often prone to learn the spurious correlations between target classes and bias attributes, like gender and race, inherent in a major portion of training data (bias-aligned samples), thus showing unfair behavior and arising controversy in the modern pluralistic and egalitarian society. In this paper, we propose a novel marginal debiased network (MDN) to learn debiased representations. More specifically, a marginal softmax loss (MSL) is designed by introducing the idea of margin penalty into the fairness problem, which assigns a larger margin for bias-conflicting samples (data without spurious correlations) than for bias-aligned ones, so as to deemphasize the spurious correlations and improve generalization on unbiased test criteria. To determine the margins, our MDN is optimized through a meta learning framework. We propose a meta equalized loss (MEL) to perceive the model fairness, and adaptively update the margin parameters by metaoptimization which requires the trained model guided by the optimal margins should minimize MEL computed on an unbiased meta-validation set. Extensive experiments on BiasedMNIST, Corrupted CIFAR-10, CelebA and UTK-Face datasets demonstrate that our MDN can achieve a remarkable performance on under-represented samples and obtain superior debiased results against the previous approaches.
Human-in-the-Loop Policy Optimization for Preference-Based Multi-Objective Reinforcement Learning
Authors: Authors: Ke Li, Han Guo
Subjects: Neural and Evolutionary Computing (cs.NE)
Abstract
Multi-objective reinforcement learning (MORL) aims to find a set of high-performing and diverse policies that address trade-offs between multiple conflicting objectives. However, in practice, decision makers (DMs) often deploy only one or a limited number of trade-off policies. Providing too many diversified trade-off policies to the DM not only significantly increases their workload but also introduces noise in multi-criterion decision-making. With this in mind, we propose a human-in-the-loop policy optimization framework for preference-based MORL that interactively identifies policies of interest. Our method proactively learns the DM's implicit preference information without requiring any a priori knowledge, which is often unavailable in real-world black-box decision scenarios. The learned preference information is used to progressively guide policy optimization towards policies of interest. We evaluate our approach against three conventional MORL algorithms that do not consider preference information and four state-of-the-art preference-based MORL algorithms on two MORL environments for robot control and smart grid management. Experimental results fully demonstrate the effectiveness of our proposed method in comparison to the other peer algorithms.
Enhancing RAW-to-sRGB with Decoupled Style Structure in Fourier Domain
Authors: Authors: Xuanhua He, Tao Hu, Guoli Wang, Zejin Wang, Run Wang, Qian Zhang, Keyu Yan, Ziyi Chen, Rui Li, Chenjun Xie, Jie Zhang, Man Zhou
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
RAW to sRGB mapping, which aims to convert RAW images from smartphones into RGB form equivalent to that of Digital Single-Lens Reflex (DSLR) cameras, has become an important area of research. However, current methods often ignore the difference between cell phone RAW images and DSLR camera RGB images, a difference that goes beyond the color matrix and extends to spatial structure due to resolution variations. Recent methods directly rebuild color mapping and spatial structure via shared deep representation, limiting optimal performance. Inspired by Image Signal Processing (ISP) pipeline, which distinguishes image restoration and enhancement, we present a novel Neural ISP framework, named FourierISP. This approach breaks the image down into style and structure within the frequency domain, allowing for independent optimization. FourierISP is comprised of three subnetworks: Phase Enhance Subnet for structural refinement, Amplitude Refine Subnet for color learning, and Color Adaptation Subnet for blending them in a smooth manner. This approach sharpens both color and structure, and extensive evaluations across varied datasets confirm that our approach realizes state-of-the-art results. Code will be available at ~\url{https://github.com/alexhe101/FourierISP}.
Trajectory-Oriented Policy Optimization with Sparse Rewards
Abstract
Deep reinforcement learning (DRL) remains challenging in tasks with sparse rewards. These sparse rewards often only indicate whether the task is partially or fully completed, meaning that many exploration actions must be performed before the agent obtains useful feedback. Hence, most existing DRL algorithms fail to learn feasible policies within a reasonable time frame. To overcome this problem, we develop an approach that exploits offline demonstration trajectories for faster and more efficient online RL in sparse reward settings. Our key insight is that by regarding offline demonstration trajectories as guidance, instead of imitating them, our method learns a policy whose state-action visitation marginal distribution matches that of offline demonstrations. Specifically, we introduce a novel trajectory distance based on maximum mean discrepancy (MMD) and formulate policy optimization as a distance-constrained optimization problem. Then, we show that this distance-constrained optimization problem can be reduced into a policy-gradient algorithm with shaped rewards learned from offline demonstrations. The proposed algorithm is evaluated on extensive discrete and continuous control tasks with sparse and deceptive rewards. The experimental results indicate that our proposed algorithm is significantly better than the baseline methods regarding diverse exploration and learning the optimal policy.
GridFormer: Point-Grid Transformer for Surface Reconstruction
Authors: Authors: Shengtao Li, Ge Gao, Yudong Liu, Yu-Shen Liu, Ming Gu
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Implicit neural networks have emerged as a crucial technology in 3D surface reconstruction. To reconstruct continuous surfaces from discrete point clouds, encoding the input points into regular grid features (plane or volume) has been commonly employed in existing approaches. However, these methods typically use the grid as an index for uniformly scattering point features. Compared with the irregular point features, the regular grid features may sacrifice some reconstruction details but improve efficiency. To take full advantage of these two types of features, we introduce a novel and high-efficiency attention mechanism between the grid and point features named Point-Grid Transformer (GridFormer). This mechanism treats the grid as a transfer point connecting the space and point cloud. Our method maximizes the spatial expressiveness of grid features and maintains computational efficiency. Furthermore, optimizing predictions over the entire space could potentially result in blurred boundaries. To address this issue, we further propose a boundary optimization strategy incorporating margin binary cross-entropy loss and boundary sampling. This approach enables us to achieve a more precise representation of the object structure. Our experiments validate that our method is effective and outperforms the state-of-the-art approaches under widely used benchmarks by producing more precise geometry reconstructions. The code is available at https://github.com/list17/GridFormer.
Training Single-Layer Morphological Perceptron Using Convex-Concave Programming
Authors: Authors: Iara Cunha, Marcos Eduardo Valle
Abstract
This paper concerns the training of a single-layer morphological perceptron using disciplined convex-concave programming (DCCP). We introduce an algorithm referred to as K-DDCCP, which combines the existing single-layer morphological perceptron (SLMP) model proposed by Ritter and Urcid with the weighted disciplined convex-concave programming (WDCCP) algorithm by Charisopoulos and Maragos. The proposed training algorithm leverages the disciplined convex-concave procedure (DCCP) and formulates a non-convex optimization problem for binary classification. To tackle this problem, the constraints are expressed as differences of convex functions, enabling the application of the DCCP package. The experimental results confirm the effectiveness of the K-DDCCP algorithm in solving binary classification problems. Overall, this work contributes to the field of morphological neural networks by proposing an algorithm that extends the capabilities of the SLMP model.
Detection and Discovery of Misinformation Sources using Attributed Webgraphs
Authors: Authors: Peter Carragher, Evan M. Williams, Kathleen M. Carley
Subjects: Social and Information Networks (cs.SI); Computers and Society (cs.CY)
Abstract
Website reliability labels underpin almost all research in misinformation detection. However, misinformation sources often exhibit transient behavior, which makes many such labeled lists obsolete over time. We demonstrate that Search Engine Optimization (SEO) attributes provide strong signals for predicting news site reliability. We introduce a novel attributed webgraph dataset with labeled news domains and their connections to outlinking and backlinking domains. We demonstrate the success of graph neural networks in detecting news site reliability using these attributed webgraphs, and show that our baseline news site reliability classifier outperforms current SoTA methods on the PoliticalNews dataset, achieving an F1 score of 0.96. Finally, we introduce and evaluate a novel graph-based algorithm for discovering previously unknown misinformation news sources.
Keyword: adam
There is no result
Keyword: gradient
A Robust Adversary Detection-Deactivation Method for Metaverse-oriented Collaborative Deep Learning
Abstract
Metaverse is trending to create a digital circumstance that can transfer the real world to an online platform supported by large quantities of real-time interactions. Pre-trained Artificial Intelligence (AI) models are demonstrating their increasing capability in aiding the metaverse to achieve an excellent response with negligible delay, and nowadays, many large models are collaboratively trained by various participants in a manner named collaborative deep learning (CDL). However, several security weaknesses can threaten the safety of the CDL training process, which might result in fatal attacks to either the pre-trained large model or the local sensitive data sets possessed by an individual entity. In CDL, malicious participants can hide within the major innocent and silently uploads deceptive parameters to degenerate the model performance, or they can abuse the downloaded parameters to construct a Generative Adversarial Network (GAN) to acquire the private information of others illegally. To compensate for these vulnerabilities, this paper proposes an adversary detection-deactivation method, which can limit and isolate the access of potential malicious participants, quarantine and disable the GAN-attack or harmful backpropagation of received threatening gradients. A detailed protection analysis has been conducted on a Multiview CDL case, and results show that the protocol can effectively prevent harmful access by heuristic manner analysis and can protect the existing model by swiftly checking received gradients using only one low-cost branch with an embedded firewall.
Shrinking Your TimeStep: Towards Low-Latency Neuromorphic Object Recognition with Spiking Neural Network
Abstract
Neuromorphic object recognition with spiking neural networks (SNNs) is the cornerstone of low-power neuromorphic computing. However, existing SNNs suffer from significant latency, utilizing 10 to 40 timesteps or more, to recognize neuromorphic objects. At low latencies, the performance of existing SNNs is drastically degraded. In this work, we propose the Shrinking SNN (SSNN) to achieve low-latency neuromorphic object recognition without reducing performance. Concretely, we alleviate the temporal redundancy in SNNs by dividing SNNs into multiple stages with progressively shrinking timesteps, which significantly reduces the inference latency. During timestep shrinkage, the temporal transformer smoothly transforms the temporal scale and preserves the information maximally. Moreover, we add multiple early classifiers to the SNN during training to mitigate the mismatch between the surrogate gradient and the true gradient, as well as the gradient vanishing/exploding, thus eliminating the performance degradation at low latency. Extensive experiments on neuromorphic datasets, CIFAR10-DVS, N-Caltech101, and DVS-Gesture have revealed that SSNN is able to improve the baseline accuracy by 6.55% ~ 21.41%. With only 5 average timesteps and without any data augmentation, SSNN is able to achieve an accuracy of 73.63% on CIFAR10-DVS. This work presents a heterogeneous temporal scale SNN and provides valuable insights into the development of high-performance, low-latency SNNs.
Spatiotemporal Monitoring of Epidemics via Solution of a Coefficient Inverse Problem
Authors: Authors: Michael V. Klibanov, Jingzhi Li, Zhipeng Yang
Abstract
Let S,I and R be susceptible, infected and recovered populations in a city affected by an epidemic. The SIR model of Lee, Liu, Tembine, Li and Osher, \emph{SIAM J. Appl. Math.},~81, 190--207, 2021 of the spatiotemoral spread of epidemics is considered. This model consists of a system of three nonlinear coupled parabolic Partial Differential Equations with respect to the space and time dependent functions S,I and R. For the first time, a Coefficient Inverse Problem (CIP) for this system is posed. The so-called \textquotedblleft convexification" numerical method for this inverse problem is constructed. The presence of the Carleman Weight Function (CWF) in the resulting regularization functional ensures the global convergence of the gradient descent method of the minimization of this functional to the true solution of the CIP, as long as the noise level tends to zero. The CWF is the function, which is used as the weight in the Carleman estimate for the corresponding Partial Differential Operator. Numerical studies demonstrate an accurate reconstruction of unknown coefficients as well as S,I,R functions inside of that city. As a by-product, uniqueness theorem for this CIP is proven. Since the minimal measured input data are required, then the proposed methodology has a potential of a significant decrease of the cost of monitoring of epidemics.
Trajectory-Oriented Policy Optimization with Sparse Rewards
Abstract
Deep reinforcement learning (DRL) remains challenging in tasks with sparse rewards. These sparse rewards often only indicate whether the task is partially or fully completed, meaning that many exploration actions must be performed before the agent obtains useful feedback. Hence, most existing DRL algorithms fail to learn feasible policies within a reasonable time frame. To overcome this problem, we develop an approach that exploits offline demonstration trajectories for faster and more efficient online RL in sparse reward settings. Our key insight is that by regarding offline demonstration trajectories as guidance, instead of imitating them, our method learns a policy whose state-action visitation marginal distribution matches that of offline demonstrations. Specifically, we introduce a novel trajectory distance based on maximum mean discrepancy (MMD) and formulate policy optimization as a distance-constrained optimization problem. Then, we show that this distance-constrained optimization problem can be reduced into a policy-gradient algorithm with shaped rewards learned from offline demonstrations. The proposed algorithm is evaluated on extensive discrete and continuous control tasks with sparse and deceptive rewards. The experimental results indicate that our proposed algorithm is significantly better than the baseline methods regarding diverse exploration and learning the optimal policy.
REDriver: Runtime Enforcement for Autonomous Vehicles
Authors: Authors: Yang Sun, Christopher M. Poskitt, Xiaodong Zhang, Jun Sun
Abstract
Autonomous driving systems (ADSs) integrate sensing, perception, drive control, and several other critical tasks in autonomous vehicles, motivating research into techniques for assessing their safety. While there are several approaches for testing and analysing them in high-fidelity simulators, ADSs may still encounter additional critical scenarios beyond those covered once they are deployed on real roads. An additional level of confidence can be established by monitoring and enforcing critical properties when the ADS is running. Existing work, however, is only able to monitor simple safety properties (e.g., avoidance of collisions) and is limited to blunt enforcement mechanisms such as hitting the emergency brakes. In this work, we propose REDriver, a general and modular approach to runtime enforcement, in which users can specify a broad range of properties (e.g., national traffic laws) in a specification language based on signal temporal logic (STL). REDriver monitors the planned trajectory of the ADS based on a quantitative semantics of STL, and uses a gradient-driven algorithm to repair the trajectory when a violation of the specification is likely. We implemented REDriver for two versions of Apollo (i.e., a popular ADS), and subjected it to a benchmark of violations of Chinese traffic laws. The results show that REDriver significantly improves Apollo's conformance to the specification with minimal overhead.
Byzantine-Resilient Gradient Coding through Local Gradient Computations
Authors: Authors: Christoph Hofmeister, Luis Maßny, Eitan Yaakobi, Rawad Bitar
Abstract
We consider gradient coding in the presence of an adversary controlling so-called malicious workers trying to corrupt the computations. Previous works propose the use of MDS codes to treat the responses from malicious workers as errors and correct them using the error-correction properties of the code. This comes at the expense of increasing the replication, i.e., the number of workers each partial gradient is computed by. In this work, we propose a way to reduce the replication to $s+1$ instead of $2s+1$ in the presence of $s$ malicious workers. Our method detects erroneous inputs from the malicious workers, transforming them into erasures. This comes at the expense of $s$ additional local computations at the main node and additional rounds of light communication between the main node and the workers. We define a general framework and give fundamental limits for fractional repetition data allocations. Our scheme is optimal in terms of replication and local computation and incurs a communication cost that is asymptotically, in the size of the dataset, a multiplicative factor away from the derived bound. We furthermore show how additional redundancy can be exploited to reduce the number of local computations and communication cost, or, alternatively, tolerate straggling workers.
Keyword: sgd
There is no result
Keyword: optimization
A Mechanistic Understanding of Alignment Algorithms: A Case Study on DPO and Toxicity
Beyond Regrets: Geometric Metrics for Bayesian Optimization
Decentralized Multi-Task Online Convex Optimization Under Random Link Failures
Fast & Fair: Efficient Second-Order Robust Optimization for Fairness in Machine Learning
From Function to Distribution Modeling: A PAC-Generative Approach to Offline Optimization
offline" data examples. While recent years have seen a flurry of work on applying various machine learning techniques to the offline optimization problem, the majority of these work focused on learning a surrogate of the unknown objective function and then applying existing optimization algorithms. While the idea of modeling the unknown objective function is intuitive and appealing, from the learning point of view it also makes it very difficult to tune the objective of the learner according to the objective of optimization. Instead of learning and then optimizing the unknown objective function, in this paper we take on a less intuitive but more direct view that optimization can be thought of as a process of sampling from a generative model. To learn an effective generative model from the offline data examples, we consider the standard technique ofre-weighting", and our main technical contribution is a probably approximately correct (PAC) lower bound on the natural optimization objective, which allows us to jointly learn a weight function and a score-based generative model. The robustly competitive performance of the proposed approach is demonstrated via empirical studies using the standard offline optimization benchmarks.Understanding LLMs: A Comprehensive Overview from Training to Inference
Joint Beamforming and Offloading Design for Integrated Sensing, Communication and Computation System
ICE-GRT: Instruction Context Enhancement by Generative Reinforcement based Transformers
Performance Trade-off and Joint Waveform Design for MIMO-OFDM DFRC Systems
View-based Explanations for Graph Neural Networks
PEFT for Speech: Unveiling Optimal Placement, Merging Strategies, and Ensemble Techniques
Marginal Debiased Network for Fair Visual Recognition
Human-in-the-Loop Policy Optimization for Preference-Based Multi-Objective Reinforcement Learning
Enhancing RAW-to-sRGB with Decoupled Style Structure in Fourier Domain
Trajectory-Oriented Policy Optimization with Sparse Rewards
GridFormer: Point-Grid Transformer for Surface Reconstruction
Training Single-Layer Morphological Perceptron Using Convex-Concave Programming
Detection and Discovery of Misinformation Sources using Attributed Webgraphs
Keyword: adam
There is no result
Keyword: gradient
A Robust Adversary Detection-Deactivation Method for Metaverse-oriented Collaborative Deep Learning
Shrinking Your TimeStep: Towards Low-Latency Neuromorphic Object Recognition with Spiking Neural Network
Spatiotemporal Monitoring of Epidemics via Solution of a Coefficient Inverse Problem
Trajectory-Oriented Policy Optimization with Sparse Rewards
REDriver: Runtime Enforcement for Autonomous Vehicles
Byzantine-Resilient Gradient Coding through Local Gradient Computations
Keyword: super-resolution
There is no result