检索结果-内蒙古大学图书馆

A formal model for multiagent Q-learning on graphs

Science China(Information Sciences) 2025年

作者： Jinzhuo LIU Guangchen JIANG Chen CHU Yong LI Zhen WANG Shuyue HU School of Software Yunnan University School of Artificial Intelligence Optics and Electronics (iO PEN)Northwestern Polytechnical University Department of Statistics School of Statistics and MathematicsYunnan University of Finance and Economics Yunnan Key Laboratory of Service Computing Yunnan University of Finance and Economics School of Information Science&Engineering Yunnan University School of Cybersecurity Northwestern Polytechnical University Shanghai Artificial Intelligence Laboratory

Understanding the dynamics of multi-agent learning has long been an important research topic. Existing research has focused mostly on 2-agent games or well-mixed populations. However, in real-world multi-agent systems, agents often interact in spatially or socially structured networks(or graphs). In this paper, we examine the dynamics of multi-agent Q-learning on graphs. Combining mean-field theory and combinatorics analysis, we present a new analytical approach to formally describe the time evolution of Q-values in the system with a topological structure. Through extensive numerical simulations, we show that our theory consistently provides an accurate depiction of the Q-learning dynamics across different typical games, initial conditions, and various graph structures, encompassing regular graphs, scale-free graphs, and random graphs. Moreover, we show that when comparing regular graphs to other types of graphs with the same average degree, the differences in the system evolution are largely attributed to the behaviors and Q-values of agents with lower degrees.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Web Service Classification Method Based on Graph Neural Network Knowledge Distillation 19

A Web Service Classification Method Based on Graph Neural Ne...

引用

2022 IEEE SmartWorld, 19th IEEE International Conference on Ubiquitous Intelligence and computing, 2022 IEEE International Conference on Autonomous and Trusted Vehicles Conference, 22nd IEEE International Conference on Scalable computing and Communications, 2022 IEEE International Conference on Digital Twin, 8th IEEE International Conference on Privacy computing and 2022 IEEE International Conference on Metaverse, SmartWorld/UIC/ATC/ScalCom/DigitalTwin/PriComp/Metaverse 2022

作者： Huang, Hao Cao, Buqing Liu, Shanpeng Zhou, Dong Tang, Mingdong Xiao, Feng Hunan University of Science and Technology School of Computer Science and Engineering Hunan Key Laboratory of Service Computing and New Software Service Technology Xiangtan China Guangdong University of Foreign Studies School of Information Science and Technology Guangzhou China

ISBN: (纸本)9798350346558

With the increasing number of Web services, service discovery becomes a significant challenge, and service classification methods based on Graph Neural Network (GNN) are considered to be an efficient way for service discovery. However, these methods rarely exploit the structure-based and feature-based prior knowledge of services. To address this problem, this paper proposes a Web service classification method based on graph neural network knowledge distillation. In this method, firstly, it employs Word2Vec to extract the semantic feature vector from service descriptions, and then builds the service relationship network based on the shared tag annotation relationship of service. Secondly, a graph neural network knowledge distillation framework for Web service classification consisting of a teacher model and a student model is designed. The teacher model extracts the knowledge of GNN model, and the student model obtains the structure-based and feature-based prior knowledge of the service relationship network from teacher model by neighbor feature aggregation and feature transformation. Finally, GNN models (GCN, GAT, SAGE, APPNP, SGC, and GCNII) as teacher models for accuracy comparison are used in the experiment. From the experiment result, the classification accuracy of the student model is 11.64%23.10% higher than that of the teacher model. © 2022 IEEE.

关键词： Web services

来源：评论

学校读者我要写书评

暂无评论

engineering the Interaction of a Humanoid Robot Pepper with Post-Stroke Patients during Training Tasks 21

Engineering the Interaction of a Humanoid Robot Pepper with ...

引用

13th ACM SIGCHI Symposium on engineering Interactive computing Systems, EICS 2021

作者： Forbrig, Peter Bundea, Alexandru Bader, Sebastian Dept. of Software Engineering University of Rostock Germany Dept. of CS/Chair of Software Engineering University of Rostock Germany Institute of Visual Chair in Mobile Multimedia Information Systems University of Rostock Institute of Visual and Analytic Computing Germany

ISBN: (纸本)9781450384490

engineering the dialog of a humanoid robot like Pepper with human beings is challenging. The software from Softbanks allows the programming of spoken sentences and the understanding of specific words for apps on the robot. However, we recognized that the fexibility of the software is limited in this way. Therefore, we developed tools based on state machines that run on a server and communicate with the robot. In this way Pepper becomes a kind of thin robot. The paper discusses the architecture of the software, its advantages and the tools used within our project E-BRAiN (Evidence-Based Robot Assistance in Neurorehabilitation). Additionally, one of our identified interaction patterns is discussed. © 2021 ACM.

关键词： Anthropomorphic robots

来源：评论

学校读者我要写书评

暂无评论

The Construction of Knowledge Graph for Personalized Online Teaching 1

引用

6th International Conference on Data Mining and Big Data, DMBD 2021

作者： Gong, Zhaokun Yu, Xiaomei Fu, Wenxiang Che, Xueyu Mao, Qian Zheng, Xiangwei School of Information Science and Engineering Shandong Normal University Jinan China Shandong Provincial Key Laboratory of Distributed Computing Software Jinan China

ISBN: (数字)9789811675027

ISBN: (纸本)9789811675010

The knowledge graph (KG) is widely used in various fields recently, while there are few achievements of knowledge graph related to educational applications. In this paper, some subject knowledge graphs are constructed to adapt to the educational development in big data environment, and the relationships between knowledge nodes are analyzed in order to provide students with a personalized online teaching for knowledge learning. Specially, a knowledge graph on python subject is constructed based on the neo4j graph database for students in programing courses. With three applications achieved on the knowledge graphs, the experimental results show that the knowledge points to students is described more clearly, and the logical relationships between the knowledge points are inferred according to the queries from students. It is verified that the knowledge graph on course learning performs significant effectiveness and efficiency in personalized online teaching. © 2021, Springer Nature Singapore Pte Ltd.

关键词： Knowledge graph

来源：评论

学校读者我要写书评

暂无评论

As Pseudo-label Free As Possible: Leveraging Adaptive Feature Generation for Sparsely Annotated Object Detection 39

As Pseudo-label Free As Possible: Leveraging Adaptive Featur...

引用

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Yao, Shuilian Liu, Yu Jia, Qi Chen, Sihong Zhuo, Wei School of Software Technology DUT-RU International School of Information Science & Engineering Dalian University of Technology Dalian China AILab Tencent Shenzhen China National Engineering Laboratory for Big Data System Computing Technology Shenzhen University Shenzhen China

ISBN: (纸本)157735897X

Compared to fully supervised object detection, training with sparse annotations typically leads to a decline in performance due to insufficient feature diversity. Existing sparsely annotated object detection (SAOD) methods often rely on pseudo-labeling strategies, but these pseudo-labels tend to introduce noise under extreme sparsity. To simultaneously avoid the impact of pseudo-label noise and enhance feature diversity, we propose a novel Adaptive Feature Generation (AdaptFG) model that generates features based on class names. This model integrates a pre-trained CLIP into a VAE-based feature generator, with its core innovation being an Adaptor that adaptively maps CLIP semantic embeddings to the object detector domain. Additionally, we introduce inter-class relationship reasoning in detector, which effectively mitigates misclassifications stemming from similar features. Extensive experimental results demonstrate that AdaptFG consistently outperforms state-of-the-art SAOD methods on the PASCAL VOC and MS COCO benchmarks. Copyright © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Learning Transferable Negative Prompts for Out-of-Distribution Detection

Learning Transferable Negative Prompts for Out-of-Distributi...

引用

Conference on Computer Vision and Pattern Recognition (CVPR)

作者： Tianqi Li Guansong Pang Xiao Bai Wenjun Miao Jin Zheng State Key Laboratory of Complex & Critical Software Environment School of Computer Science and Engineering Jiangxi Research Institute Beihang University China School of Computing and Information Systems Singapore Management University

ISBN: (数字)9798350353006

ISBN: (纸本)9798350353013

Existing prompt learning methods have shown certain capabilities in Out-of-Distribution (OOD) detection, but the lack of OOD images in the target dataset in their training can lead to mismatches between OOD images and In-Distribution (ID) categories, resulting in a high false positive rate. To address this issue, we introduce a novel OOD detection method, named ‘NegPrompt’, to learn a set of negative prompts, each representing a negative connotation of a given class label, for delineating the boundaries between ID and OOD images. It learns such negative prompts with ID data only, without any reliance on external out-lier data. Further, current methods assume the availability of samples of all ID classes, rendering them ineffective in open-vocabulary learning scenarios where the inference stage can contain novel ID classes not present during training. In contrast, our learned negative prompts are transferable to novel class labels. Experiments on various ImageNet benchmarks show that NegPrompt surpasses state-of-the-art prompt-learning-based OOD detection methods and maintains a consistent lead in hard OOD detection in closed- and open-vocabulary classification scenarios. Code is available at https://***/mala-lab/negprompt.

关键词： Training Learning systems Computer vision Codes Computational modeling Semantics Lead

来源：评论

学校读者我要写书评

暂无评论

Optimal Bounds for Distinct Quartics

arXiv

引用

arXiv 2024年

作者： Charalampopoulos, Panagiotis Gawrychowski, Pawel Ghazawi, Samah School of Computing and Mathematical Sciences Birkbeck University of London United Kingdom Institute of Computer Science University of Wroclaw Poland Department of Software Engineering Braude College of Engineering Karmiel Israel Department of Computer Science University of Haifa Israel

A fundamental concept related to strings is that of repetitions. It has been extensively studied in many versions, from both purely combinatorial and algorithmic angles. One of the most basic questions is how many distinct squares, i.e., distinct strings of the form UU, a string of length n can contain as fragments. It turns out that this is always O(n), and the bound cannot be improved to sublinear in n [Fraenkel and Simpson, JCTA 1998]. Several similar questions about repetitions in strings have been considered, and by now we seem to have a good understanding of their repetitive structure. For higher-dimensional strings, the basic concept of periodicity has been successfully extended and applied to design efficient algorithms—it is inherently more complex than for regular strings. Extending the notion of repetitions and understanding the repetitive structure of higher-dimensional strings is however far from complete. Quartics were introduced by Apostolico and Brimkov [TCS 2000] as analogues of squares in two dimensions. Charalampopoulos, Radoszewski, Rytter, Waleń, and Zuba [ESA 2020] proved that the number of distinct quartics in an n × n 2D string is O(n2 log2 n) and that they can be computed in O(n2 log2 n) time. Gawrychowski, Ghazawi, and Landau [SPIRE 2021] constructed an infinite family of n×n 2D strings with Ω(n2 log n) distinct quartics. This brings the challenge of determining asymptotically tight bounds. Here, we settle both the combinatorial and the algorithmic aspects of this question: the number of distinct quartics in an n × n 2D string is O(n2 log n) and they can be computed in the worst-case optimal O(n2 log n) time. As expected, our solution heavily exploits the periodic structure implied by occurrences of quartics. However, the two-dimensional nature of the problem introduces some technical challenges. Somewhat surprisingly, we overcome the final challenge for the combinatorial bound using a result of Marcus and Tardos [JCTA 2004] for permutation

关键词： Optimization

来源：评论

学校读者我要写书评

暂无评论

Deep Hierarchy Quantization Compression algorithm based on Dynamic Sampling

arXiv

引用

arXiv 2022年

作者： Jiang, Wan Liu, Gang Chen, Xiaofeng Zhou, Yipeng College of Computer Science and Software Engineering ShenZhen University China Department of Computing Macquarie University Australia

Unlike traditional distributed machine learning, federated learning stores data locally for training and then aggregates the models on the server, which solves the data security problem that may arise in traditional distributed machine learning. However, during the training process, the transmission of model parameters can impose a significant load on the network bandwidth. It has been pointed out that the vast majority of model parameters are redundant during model parameter transmission. In this paper, we explore the data distribution law of selected partial model parameters on this basis, and propose a deep hierarchical quantization compression algorithm, which further compresses the model and reduces the network load brought by data transmission through the hierarchical quantization of model parameters. And we adopt a dynamic sampling strategy for the selection of clients to accelerate the convergence of the model. Experimental results on different public datasets demonstrate the effectiveness of our algorithm. Copyright © 2022, The Authors. All rights reserved.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

Existence of strongly nonlocal sets of three states in any N-partite system

arXiv

引用

arXiv 2024年

作者： Xiong, Zong-Xing Li, Mao-Sheng Institute of Quantum Computing and Software School of Computer Science and Engineering Sun Yat-Sen University Guangzhou510006 China School of Mathematics South China University of Technology Guangzhou510641 China

The notion of strong nonlocality, which refers to local irreducibility of a set of orthogonal multipartite quantum states across each bipartition of the subsystems, was put forward by Halder et al. in [Phys. Rev. Lett. 122, 040403 (2019)]. Here, we show the existence of three orthogonal quantum states in (C2)⊗N that cannot be perfectly distinguished locally across any bipartition of the subsystems. Specifically, all these three states are genuinely entangled, among which two are the N-qubit GHZ pairs. Since any three locally indistinguishable states are always locally irreducible, the three N-partite orthogonal states we present are strongly nonlocal. Thus, the caridnality of strongly nonlocal sets here is dramatically smaller than all known ones. © 2024, CC BY.

关键词： Quantum entanglement

来源：评论

学校读者我要写书评

暂无评论

Oxygen: A Distributed Health Care Framework for Patient Health Record Management and Pharmaceutical Diagnosis

Oxygen: A Distributed Health Care Framework for Patient Heal...

引用

Advancements in computing (ICAC), International Conference on

作者： Maleesha Wickramarathna Kithmini De Silva Vihanga Lekamalage Janith Senanayake Jeewaka Perera Laneesha Ruggahakotuwa Department of Software Engineering Faculty of Computing Sri Lanka Institute of Information Technology Malabe Sri Lanka

ISBN: (纸本)9798350398106

With the COVID-19 pandemic, the world is confronting various healthcare issues, and healthcare automation is more crucial than ever. The pandemic has revealed the limitations of existing digital healthcare systems to manage public health emergencies. There is no registered population for many healthcare institutions in Sri Lanka, as a result, there is a communication gap. Electronic Health Record systems (EHRs) are becoming popular to share patient details but accessing scattered data across several EHRs while safeguarding patient privacy remains a challenge. Most of these medical records are in printed format and manually entering those into EHR systems is time-consuming and error prone. Not only that pharmaceutical error is a critical healthcare problem, but it is even riskier to visit doctors for pharmaceutical diagnosis during a pandemic. This research introduces a Blockchain-based patient health record system, an Optical Character Recognition (OCR) and Natural Language Processing (NLP) based Medical Document Scanner, a Drug Identifier based on Image Processing and a Medical Chatbot powered by NLP as four novel approaches to address these issues. Altogether with the results, this research aims at introducing a solution for the limitations in healthcare while providing a distributed healthcare framework for the healthcare community worldwide.

关键词： Oxygen Pandemics Image processing Optical character recognition Sociology Medical services Statistics

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：