检索结果-内蒙古大学图书馆

Prediction Guided Meta-Learning for Multi-Objective Reinforcement Learning

学校读者我要写书评

暂无评论

Prediction Guided Meta-Learning for Multi-Objective Reinforc...

Congress on Evolutionary Computation

作者： Fei-Yu Liu Chao Qian School of Computer Science and Technology University of Science and Technology of China Hefei China State Key Laboratory for Novel Software Technology Nanjing University Nanjing China

Many real-world control problems consist of several different, possibly conflicting, objectives, which require finding a high-quality set of policies that are optimal for different objective preferences. Extensive research mainly focused on how to obtain a high-quality approximated Pareto set of policies, while another important research direction studies how to adapt to new objective preferences quickly. In this paper, we propose a new multi-objective reinforcement learning (MORL) algorithm so-called PG-Meta-MORL for achieving both goals. PG-Meta-MORL frames MORL as a meta-learning problem and iteratively optimizes a meta-policy using multiple tasks with objective preferences selected based on a prediction model, which is trained to guide the optimization process towards best improving the quality of the current Pareto set of policies. The empirical results on several multi-objective continuous control problems show that PG-Meta-MORL can find a high-quality approximated Pareto set of policies, and meanwhile, the obtained meta-policy can be adapted well to new objective preferences using few-shot interactions with the environment.

关键词： Adaptation models Reinforcement learning Evolutionary computation Predictive models Prediction algorithms Approximation algorithms Task analysis

Implicit Multi-Scale Swin Transformer Network for Image Denoising

学校读者我要写书评

暂无评论

IEEE Transactions on Consumer Electronics 2025年

作者： Zhang, Qi Ding, Yuwei Zhang, Weiqi Zhu, Yian Zhang, Bob Lin, Jerry Chun-Wei Harbin Institute of Technology at Weihai School of Economics and Management Weihai China Nanjing University State Key Laboratory for Novel Software Technology Nanjing China Northwestern Polytechnical University School of Computer Science Xi’an China Northwestern Polytechnical University Yangtze River Delta Research Institute Taicang China University of Macau PAMI Research Group Department of Computer and Information Science Taipa China Department of Distributed Systems and IT Devices Electrical Engineering and Mathematical Sciences Silesian university of technology Poland

Image denoising has been used in various edge computing scenarios such as consumer electronics to improve the image quality and user experience. Existing image denoising methods based on Convolutional Neural Networks (CNNs) and vision Transformers achieve good performance by empirically utilizing residual neural network (ResNet) as basic component. However, ResNet lacks interpretability in network design and also long-term memory of features, potentially restricting the performance of denoising networks. In this paper, we propose an Implicit Multi-scale Swin Transformer Network (IMSNet) for image denoising, which introduces the implicit Euler scheme from the feature memory perspective. Specifically, densely connected implicit feature extraction blocks (IFEBs) are designed to learn the residual mapping between noisy and clean images. The IFEB reformulates the initial skip connection of ResNet based on the implicit Euler discretization, providing both network interpretability and long-term memory. In IFEB, the multi-scale swin Transformer block (MSB) is designed as the implicit layer to capture spatial details and non-local contextual information at different scales. Additionally, a cross-layer feature fusion block (CLFF) is proposed to further improve feature reuse capabilities. Compared to existing denoising networks, the extensive experiments demonstrate the superior performance of our IMSNet in various image denoising tasks, and the flexibility in the practical applications with proposed light model on resource-restricted platforms such as consumer electronic devices. © 1975-2011 IEEE.

关键词： Image denoising

Exploring Leximin Principle for Fair Core-Selecting Combinatorial Auctions: Payment Rule Design and Implementation

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Cheng, Hao Kong, Shufeng Deng, Yanchen Liu, Caihua Wu, Xiaohu An, Bo Wang, Chongjun State Key Laboratory for Novel Software Technology Nanjing University China School of Software Engineering Sun Yat-Sen University China School of Computer Science and Engineering Nanyang Technological University Singapore School of Artificial Intelligence Guilin University of Electronic Technology China National Engineering Research Center of Mobile Network Technologies Beijing University of Posts and Telecommunications China

Core-selecting combinatorial auctions (CAs) restrict the auction result in the core such that no coalitions could improve their utilities by engaging in collusion. The minimum-revenue-core (MRC) rule is a widely used core-selecting payment rule to maximize the total utilities of all bidders. However, the MRC rule can suffer from severe unfairness since it ignores individuals’ utilities. To address this limitation, we propose to explore the leximin principle to achieve fairness in core-selecting CAs since the leximin principle prefers to maximize the utility of the worst-off;the resulting bidder-leximin-optimal (BLO) payment rule is then theoretically analyzed and an effective algorithm is further provided to compute the BLO outcome. Moreover, we conduct extensive experiments to show that our algorithm returns fairer utility distributions and is faster than existing algorithms of core-selecting payment rules. Copyright © 2023, The Authors. All rights reserved.

关键词： Machine learning

Mutual Information Guided Backdoor Mitigation for Pre-trained Encoders

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Han, Tingxu Sun, Weisong Ding, Ziqi Fang, Chunrong Qian, Hanwei Li, Jiaxun Chen, Zhenyu Zhang, Xiangyu The State Key Laboratory for Novel Software Technology Nanjing University Nanjing210093 China The School of Computer Science and Engineering Nanyang Technological University Nanyang639798 Singapore The School of Computer Science and Engineering University of New South Wales NSW2052 Australia The School of Mathematical Sciences Soochow University Suzhou215000 China The School of Computer Sciences Purdue University West Lafayette47907 United States

Self-supervised learning (SSL) is increasingly attractive for pre-training encoders without requiring labeled data. Downstream tasks built on top of those pre-trained encoders can achieve nearly state-of-the-art performance. The pre-trained encoders by SSL, however, are vulnerable to backdoor attacks as demonstrated by existing studies. Numerous backdoor mitigation techniques are designed for downstream task models. However, their effectiveness is impaired and limited when adapted to pre-trained encoders, due to the lack of label information when pre-training. To address backdoor attacks against pre-trained encoders, in this paper, we innovatively propose a mutual information guided backdoor mitigation technique, named MIMIC. MIMIC treats the potentially backdoored encoder as the teacher net and employs knowledge distillation to distill a clean student encoder from the teacher net. Different from existing knowledge distillation approaches, MIMIC initializes the student with random weights, inheriting no backdoors from teacher nets. Then MIMIC leverages mutual information between each layer and extracted features to locate where benign knowledge lies in the teacher net, with which distillation is deployed to clone clean features from teacher to student. We craft the distillation loss with two aspects, including clone loss and attention loss, aiming to mitigate backdoors and maintain encoder performance at the same time. Our evaluation conducted on two backdoor attacks in SSL demonstrates that MIMIC can significantly reduce the attack success rate by only utilizing Copyright © 2024, The Authors. All rights reserved.

关键词： Distillation

VeriQBench: A Benchmark for Multiple Types of Quantum Circuits

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Chen, Kean Fang, Wang Guan, Ji Hong, Xin Huang, Mingyu Liu, Junyi Wang, Qisheng Ying, Mingsheng State Key Laboratory of Computer Science Institute of Software Chinese Academy of Sciences China Department of Computer Science and Technology Tsinghua University China The University of Chinese Academy of Sciences China Centre for Quantum Software and Information University of Technology Sydney Australia

In this paper, we introduce VeriQBench-an open source benchmark for quantum circuits. It offers high-level quantum circuit abstractions of various circuit types, including: 1) combinational, 2) dynamic, 3) sequen-tial, and 4) variational quantum circuits, which cover almost all existing types of quantum circuits in the literature. Meanwhile, VeriQBench is a versatile benchmark which can be used in verifying quantum software for different applications, as is evidenced by the existing works includ-ing quantum circuit verification (e.g., equivalence checking [Hon+21a;WLY21] and model checking [Yin21]), simulation (e.g., fault simulation), testing (e.g., test pattern generation [CY22]) and debugging (e.g., run-time assertions [Li+20b]). All the circuits are described in OpenQASM and are validated on Qiskit and QCOR simulators. With the hope that it can be used by other researchers, VeriQBench is released at: https:// ***/Veri-Q/Benchmark. © 2022, CC BY.

关键词： Application programs

Revisiting dependence cluster metrics based defect prediction 32

学校读者我要写书评

暂无评论

Revisiting dependence cluster metrics based defect predictio...

32nd International Conference on software Engineering and Knowledge Engineering, SEKE 2020

作者： Huang, Qiguo Chen, Xiang Li, Zhengliang Ni, Chao Gu, Qing State Key Laboratory for Novel Software Technology Nanjing University Nanjing210023 China School of Computer Science and Technology Nantong University Nantong226019 China

ISBN: (纸本)1891706500

A dependence cluster is a set of program elements that all depend upon each other. Prior empirical studies have found that the dependence cluster based metrics are useful in effort-aware defect prediction. However, it is still unknown whether they are useful in non-effort-aware defect prediction. In this paper, we perform empirical studies to investigate this issue. We use the product, process, and network metrics to build the "B" model (baseline model), and then use the product, process, network and dependence cluster metrics to build the "B+C" model. Our experimental results, based on five well-known open-source systems, show that the dependence clusters are useful for non-effort-aware defect prediction. These findings help us better understand how dependence clusters influence non-effort-aware defect prediction. © 2020 Knowledge Systems Institute Graduate School. All rights reserved.

关键词： Forecasting

Multi-SIGATnet: A multimodal schizophrenia MRI classification algorithm using sparse interaction mechanisms and graph attention networks

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Jiao, Yuhong Miao, Jiaqing Gong, Jinnan He, Hui Liang, Ping Luo, Cheng Tan, Ying The Key Laboratory for Computer Systems of State Ethnic Affairs Commission Southwest Minzu University Chengdu610041 China School of computing and software Chengdu Neusoft University Chengdu611844 China School of Computer Science Chengdu University of Information Technology Chengdu610225 China The Key Laboratory for NeuroInformation of Ministry of Education High-Field Magnetic Resonance Brain Imaging Key Laboratory of Sichuan Province Center for Information in Medicine University of Electronic Science and Technology of China Chengdu610054 China

Schizophrenia is a serious psychiatric disorder. Its pathogenesis is not completely clear, making it difficult to treat patients precisely. Because of the complicated non-Euclidean network structure of the human brain, learning critical information from brain networks remains difficult. To effectively capture the topological information of brain neural networks, a novel multimodal graph attention network based on sparse interaction mechanism (Multi-SIGATnet) was proposed for SZ classification was proposed for SZ classification. Firstly, structural and functional information were fused into multimodal data to obtain more comprehensive and abundant features for patients with SZ. Subsequently, a sparse interaction mechanism was proposed to effectively extract salient features and enhance the feature representation capability. By enhancing the strong connections and weakening the weak connections between feature information based on an asymmetric convolutional network, high-order interactive features were captured. Moreover, sparse learning strategies were designed to filter out redundant connections to improve model performance. Finally, local and global features were updated in accordance with the topological features and connection weight constraints of the higher-order brain network, the features being projected to the classification target space for disorder classification. The effectiveness of the model is verified on the Center for Biomedical Research Excellence (COBRE) and University of California Los Angeles (UCLA) datasets, achieving 81.9\% and 75.8\% average accuracy, respectively, 4.6\% and 5.5\% higher than the graph attention network (GAT) method. Experiments showed that the Multi-SIGATnet method exhibited good performance in identifying SZ. Copyright © 2024, The Authors. All rights reserved.

关键词： Convolution

Local Search For Satisfiability Modulo Integer Arithmetic Theories

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Cai, Shaowei Li, Bohan Zhang, Xindi State Key Laboratory of Computer Science Institute of Software Chinese Academy of Sciences School of Computer Science and Technology University of Chinese Academy of Sciences Beijing China

Satisfiability Modulo Theories (SMT) refers to the problem of deciding the satisfiability of a formula with respect to certain background first-order theories. In this paper, we focus on Satisfiablity Modulo Integer Arithmetic, which is referred to as SMT(IA), including both linear and non-linear integer arithmetic theories. Dominant approaches to SMT rely on calling a CDCL-based SAT solver, either in a lazy or eager flavour. Local search, a competitive approach to solving combinatorial problems including SAT, however, has not been well studied for SMT. We develop the first local-search algorithm for SMT(IA) by directly operating on variables, breaking through the traditional framework. We propose a local-search framework by considering the distinctions between Boolean and integer variables. Moreover, we design a novel operator and scoring functions tailored for integer arithmetic, as well as a two-level operation selection heuristic. Putting these together, we develop a local search SMT(IA) solver called LocalSMT. Experiments are carried out to evaluate LocalSMT on benchmark sets from SMT-LIB. The results show that LocalSMT is competitive and complementary with state-of-the-art SMT solvers, and performs particularly well on those formulae with only integer variables. A simple sequential portfolio with Z3 improves the state-of-the-art on satisfiable benchmark sets from SMT-LIB. © 2022, CC BY.

关键词： Local search (optimization)

Efficient algorithms for generalized linear bandits with heavy-tailed rewards 23

学校读者我要写书评

暂无评论

Efficient algorithms for generalized linear bandits with hea...

Proceedings of the 37th International Conference on Neural Information Processing Systems

作者： Bo Xue Yimu Wang Yuanyu Wan Jinfeng Yi Lijun Zhang Department of Computer Science City University of Hong Kong Hong Kong China and The City University of Hong Kong Shenzhen Research Institute Shenzhen China Cheriton School of Computer Science University of Waterloo Waterloo Canada School of Software Technology Zhejiang University Ningbo China JD AI Research Beijing China National Key Laboratory for Novel Software Technology Nanjing University Nanjing China and Peng Cheng Laboratory Shenzhen China

This paper investigates the problem of generalized linear bandits with heavy-tailed rewards, whose (1 + ε)-th moment is bounded for some ε ∈ (0,1]. Although there exist methods for generalized linear bandits, most of them focus on bounded or sub-Gaussian rewards and are not well-suited for many real-world scenarios, such as financial markets and web-advertising. To address this issue, we propose two novel algorithms based on truncation and mean of medians. These algorithms achieve an almost optimal regret bound of $\widetilde{O}(dT^{\frac{1}{1+\epsilon}})$, where d is the dimension of contextual information and T is the time horizon. Our truncation-based algorithm supports online learning, distinguishing it from existing truncation-based approaches. Additionally, our mean-of-medians-based algorithm requires only O(log T) rewards and one estimator per epoch, making it more practical. Moreover, our algorithms improve the regret bounds by a logarithmic factor compared to existing algorithms when ε =1. Numerical experimental results confirm the merits of our algorithms.

关键词：