检索结果-内蒙古大学图书馆

Model-Based Offline Reinforcement Learning with Adversarial Data Augmentation

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Cao, Hongye Feng, Fan Huo, Jing Yang, Shangdong Fang, Meng Yang, Tianpei Gao, Yang National Key Laboratory for Novel Software Technology Nanjing University Nanjing210093 China Department of Electrical Engineering City University of Hong Kong Hong Kong School of Computer Science Nanjing University of Posts and Telecommunications Nanjing210023 China Department of Computer Science University of Liverpool LiverpoolL69 3BX United Kingdom

Model-based offline Reinforcement Learning (RL) constructs environment models from offline datasets to perform conservative policy optimization. Existing approaches focus on learning state transitions through ensemble models, rollouting conservative estimation to mitigate extrapolation errors. However, the static data makes it challenging to develop a robust policy, and offline agents cannot access the environment to gather new data. To address these challenges, we introduce Model-based Offline Reinforcement learning with AdversariaL data augmentation (MORAL). In MORAL, we replace the fixed horizon rollout by employing adversaria data augmentation to execute alternating sampling with ensemble models to enrich training data. Specifically, this adversarial process dynamically selects ensemble models against policy for biased sampling, mitigating the optimistic estimation of fixed models, thus robustly expanding the training data for policy optimization. Moreover, a differential factor is integrated into the adversarial process for regularization, ensuring error minimization in extrapolations. This data-augmented optimization adapts to diverse offline tasks without rollout horizon tuning, showing remarkable applicability. Extensive experiments on D4RL benchmark demonstrate that MORAL outperforms other model-based offline RL methods in terms of policy learning and sample efficiency. © 2025, CC BY.

关键词： Adversarial machine learning

EFM: A Negative Network Public Opinion Early Warning Model Based on Evolutionary Feature Mining

学校读者我要写书评

暂无评论

EFM: A Negative Network Public Opinion Early Warning Model B...

IEEE International Conference on Data science in Cyberspace (DSC)

作者： Fanfeixue Meng Jingchen Wu Feifan Song Xu Wu Lan Hu Key Laboratory of Trustworthy Distributed Computing and Service Ministry of Education School of Cyberspace Security BUPT Beijing University of Posts and Telecommunications Beijing China School of Computer Science (National Pilot Software Engineering School) BUPT School of Economics and Finance Changchun Finance College

With the popularity of the Internet and the increase in the number of netizens, it is of great significance to effectively monitor and predict the negative network public opinion on social media platforms. Most mainstream negative network public opinion early warning models mainly focus on mining textual features, while neglecting the response and dissemination among netizens during the process of negative network public opinion fermentation. This is one of the important reasons for the accuracy bottleneck of them. This article proposes a negative network public opinion early warning model based on evolutionary feature mining. Firstly, by using entity linking techniques based on knowledge graphs, domain prior knowledge is supplemented for the text need to be analyzed for negative network public opinion. Then, the text sentiment features extracted based on long short-term memory networks and the evolutionary features extracted based on graph convolutional networks are fused through an enhanced-feature fusion method. Finally, the features are used for classification and early warning. Experimental results based on the Weibo public dataset show that the EFM model has higher early warning performance for negative network public opinion compared to other baseline models. The accuracy of this model is 0.959.

关键词：

Learning the Relation between Similarity Loss and Clustering Loss in Self-Supervised Learning

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Ge, Jidong Liu, Yuxiang Gui, Jie Fang, Lanting Lin, Ming Kwok, James Tin-Yau Huang, LiGuo Luo, Bin State Key Laboratory for Novel Software Technology Software Institute Nanjing University Nanjing210093 China The School of Cyber Science and Engineering Southeast University Purple Mountain Laboratories Nanjing210000 China *** L.L.C. United States Alibaba Group China The Department of Computer Science and Engineering The Hong Kong University of Science and Technology 999077 Hong Kong The Department of Computer Science Southern Methodist University DallasTX75205 United States

Self-supervised learning enables networks to learn discriminative features from massive data itself. Most state-of-the-art methods maximize the similarity between two augmentations of one image based on contrastive learning. By utilizing the consistency of two augmentations, the burden of manual annotations can be freed. Contrastive learning exploits instance-level information to learn robust features. However, the learned information is probably confined to different views of the same instance. In this paper, we attempt to leverage the similarity between two distinct images to boost representation in self-supervised learning. In contrast to instance-level information, the similarity between two distinct images may provide more useful information. Besides, we analyze the relation between similarity loss and feature-level cross-entropy loss. These two losses are essential for most deep learning methods. However, the relation between these two losses is not clear. Similarity loss helps obtain instance-level representation, while feature-level cross-entropy loss helps mine the similarity between two distinct images. We provide theoretical analyses and experiments to show that a suitable combination of these two losses can get state-of-the-art results. Code is available at https://***/guijiejie/ICCL. Copyright © 2023, The Authors. All rights reserved.

关键词： Image classification

Structure-Preserving Graph Representation Learning

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Fang, Ruiyi Wen, Liangjian Kang, Zhao Liu, Jianzhuang School of Information and Software Engineering University of Electronic Science and Technology of China China School of Computer Science and Engineering University of Electronic Science and Technology of China China Shenzhen Institutes of Advanced Technology University of Chinese Academy of Sciences China

Though graph representation learning (GRL) has made significant progress, it is still a challenge to extract and embed the rich topological structure and feature information in an adequate way. Most existing methods focus on local structure and fail to fully incorporate the global topological structure. To this end, we propose a novel Structure-Preserving Graph Representation Learning (SPGRL) method, to fully capture the structure information of graphs. Specifically, to reduce the uncertainty and misinformation of the original graph, we construct a feature graph as a complementary view via kNearest Neighbor method. The feature graph can be used to contrast at node-level to capture the local relation. Besides, we retain the global topological structure information by maximizing the mutual information (MI) of the whole graph and feature embeddings, which is theoretically reduced to exchanging the feature embeddings of the feature and the original graphs to reconstruct themselves. Extensive experiments show that our method has quite superior performance on semi-supervised node classification task and excellent robustness under noise perturbation on graph structure or node features. The source code is available at https://***/uestc-lese/SPGRL. Copyright © 2022, The Authors. All rights reserved.

关键词： Classification (of information)

A Simulation Method for Industrial Control Honeynet Based on Time Series Prediction

学校读者我要写书评

暂无评论

A Simulation Method for Industrial Control Honeynet Based on...

Information science, Parallel and Distributed Systems (ISPDS), International Conference on

作者： Tianqi Jiang Yu Yao Chao Xu Wei Yang Xiaoming Zhou Miao He Wenjie Zhang Qianyi Wang Meng Meng Rui Wang School of Computer Science and Engineering Northeastern University Shenyang China State Grid Liaoning Xintong Company Shenyang China Software College Shenyang China State Grid Liaoning Electric Power Co. Ltd. Shenyang China

ISBN: (数字)9798350352719

ISBN: (纸本)9798350352726

This paper proposes a time series prediction based industrial control honey network simulation method to address the issue that many current industrial control honey network solutions only consider the response interaction simulation of attackers within a certain period of time and cannot effectively achieve communication simulation and collaborative work between honeypots. Firstly, the deployment and basic interaction simulation of the industrial control honeynet have been achieved; Then, a communication simulation method between devices within the industrial control honey network was studied and implemented to simulate collaborative work scenarios between devices; Secondly, using time series prediction algorithms to generate simulation data for various devices within the industrial control honey network, improving the sustained response capability of the industrial control honey network; Finally, comparative experiments have shown that the method proposed in this article can improve the ability to trap attackers, provide more data for security analysis, and help users establish a comprehensive industrial control network security protection system.

关键词： Information science Protocols Adaptive systems Federated learning Industrial control Time series analysis Predictive models

A Study on the Construction and Practical Application of the Smart Education Data Platform at Guizhou Traffic Technician and Transportation College

学校读者我要写书评

暂无评论

A Study on the Construction and Practical Application of the...

IEEE International Symposium on Information (IT) in Medicine and Education, ITME

作者： Hong Li Hong Zhang Shaoxun Deng Bo Li Research and Development Department Guizhou Bit Software Co. Ltd. Guiyang China Department of Management Engineering Guizhou Traffic Technician and Transportation College Guiyang China School of Computer Science and Technology Guizhou University Guiyang China

ISBN: (数字)9798350356670

ISBN: (纸本)9798350356687

This paper investigates the digital transformation at Guizhou Traffic Technician and Transportation College, aligned with the 'China Education Modernization 2035' agenda. It introduces a microservices-based data platform aimed at upgrading educational IT systems, enhancing instructional quality, and informing decisions through smarter data use. The approach integrates microservices, cloud, big data, and AI to achieve efficient data handling and analytics, leading to improved management and decision-making processes. The study documents the platform's architecture, security, challenges, and solutions, along with successes like data visualization and automated reporting. Findings emphasize the project's relevance for vocational education's digital transition, pointing out future challenges, including data security and technology updates. The research concludes with key takeaways for institutions undergoing digitalization, underscoring the paramountcy of data security and tech optimization.

关键词： Data security Digital transformation Data handling Decision making Microservice architectures Transportation Data visualization Educational technology Information technology Optimization

ESA: Example Sieve Approach for Multi-Positive and Unlabeled Learning

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Li, Zhongnian Wei, Meng Ying, Peng Xu, Xinzheng School of Computer Science and Technology China University of Mining and Technology Xuzhou China Mine Digitization Engineering Research Center of the Ministry of Education Xuzhou China State Key Lab. for Novel Software Technology Nanjing University Nanjing China

Learning from Multi-Positive and Unlabeled (MPU) data has gradually attracted significant attention from practical applications. Unfortunately, the risk of MPU also suffer from the shift of minimum risk, particularly when the models are very flexible as shown in Fig.1. In this paper, to alleviate the shifting of minimum risk problem, we propose an Example Sieve Approach (ESA) to select examples for training a multi-class classifier. Specifically, we sieve out some examples by utilizing the Certain Loss (CL) value of each example in the training stage and analyze the consistency of the proposed risk estimator. Besides, we show that the estimation error of proposed ESA obtains the optimal parametric convergence rate. Extensive experiments on various real-world datasets show the proposed approach outperforms previous methods. Source code is available at https://***/WilsonMqz/ESA Copyright © 2024, The Authors. All rights reserved.

关键词： Contrastive Learning

Correction to: Region Centric Minutiae Propagation Measure Orient Forgery Detection with Finger Print Analysis in Health Care Systems

学校读者我要写书评

暂无评论

Neural Processing Letters 2022年第1期55卷 33-34页

作者： Baskar, M. Renuka Devi, R. Ramkumar, J. Kalyanasundaram, P. Suchithra, M. Amutha, B. Department of Computer Science and Engineering School of Computing College of Engineering and Technology SRM Institute of Science and Technology Kattankulathur Kanchipuram India Department of Software Engineering School of Computing College of Engineering and Technology SRM Institute of Science and Technology Kattankulathur Kanchipuram India Department of Wireless Sensor Networks Institute of ECE Saveetha School of Engineering Saveetha Institute of Medical and Technical Sciences Chennai India

Adaptive Dual Channel Convolution Hypergraph Representation Learning for Technological Intellectual Property

学校读者我要写书评

暂无评论

Adaptive Dual Channel Convolution Hypergraph Representation ...

IEEE International Conference on Cloud Computing and Intelligence Systems (CCIS)

作者： Yuxin Liu Yawen Li Yingxia Shao Zeli Guan Beijing Key Laboratory of Intelligent Communication Software and Multimedia School of Computer Science (National Pilot Software Engineering School) Beijing University of Posts and Telecommunications Beijing School of Economics and Management Beijing University of Posts and Telecommunications Beijing

ISBN: (纸本)9781665456579

In the age of big data, the demand for hidden information mining in technological intellectual property (tech-IP) is increasing in discrete countries. Definitely, a considerable number of graph learning algorithms for technological intellectual property have been proposed. The goal is to model the technological intellectual property entities and their relationships through the graph structure and use the neural network algorithm to extract the hidden structure information in the graph. However, most of the existing graph learning algorithms merely focus on the information mining of binary relations in technological intellectual property, ignoring the higher-order information hidden in non-binary relations. Therefore, a hypergraph neural network model based on dual channel convolution is proposed. For the hypergraph constructed from technological intellectual property data, the hypergraph channel and the line expanded graph channel of the hypergraph are used to learn the hypergraph, and the attention mechanism is introduced to adaptively fuse the output representations of the two channels. The proposed model outperforms the existing approaches on a variety of datasets.

关键词： Representation learning Adaptation models Convolution Fuses Computational modeling Neural networks Intellectual property