检索结果-内蒙古大学图书馆

Woodpecker: hallucination correction for multimodal large language models

Science China(Information Sciences) 2024年第12期67卷 52-64页

作者： Shukang YIN Chaoyou FU Sirui ZHAO Tong XU Hao WANG Dianbo SUI Yunhang SHEN Ke LI Xing SUN Enhong CHEN School of Artificial Intelligence and Data Science University of Science and Technology of China State Key Laboratory for Novel Software Technology Nanjing University School of Intelligence Science and Technology Nanjing University Institute of Automation Chinese Academy of Sciences YouTu

Hallucinations is a big shadow hanging over the rapidly evolving multimodal large language models(MLLMs), referring to that the generated text is inconsistent with the image content. To mitigate hallucinations, existing studies mainly resort to an instruction-tuning manner that requires retraining the models with specific data. In this paper, we pave a different way, introducing a training-free method named Woodpecker. Like woodpeckers heal trees, it picks out and corrects hallucinations from the generated text. Concretely, Woodpecker consists of five stages: key concept extraction, question formulation, visual knowledge validation, visual claim generation, and hallucination correction. Implemented in a post-remedy manner, Woodpecker can easily serve different MLLMs, while being interpretable by accessing intermediate outputs of the five stages. We evaluate Woodpecker both quantitatively and qualitatively and show the huge potential of this new paradigm. On the POPE benchmark, our method obtains a 30.66%/24.33% improvement in accuracy over the baseline MiniGPT-4/mPLUG-Owl. The source code is released at https://***/BradyFU/Woodpecker.

关键词： multimodal learning multimodal large language models hallucination correction large language models vision and language

来源：评论

学校读者我要写书评

暂无评论

Computational Experiments for Complex Social Systems:Experiment Design and Generative Explanation

引用

IEEE/CAA Journal of Automatica Sinica 2024年第4期11卷 1022-1038页

作者： Xiao Xue Deyu Zhou Xiangning Yu Gang Wang Juanjuan Li Xia Xie Lizhen Cui Fei-Yue Wang the School of Computer Software College of Intelligence and ComputingTianjin UniversityTianjin 300350China the School of Software and the Joint SDUNTU Centre for Artificial Intelligence Research(C-FAIR) Shandong UniversityJinan 250101China the Information School Xi’an University of Financial and EconomicsXi’an 710100China the State Key Laboratory of Multimodal Artificial Intelligence Systems Institute of AutomationChinese Academy of SciencesBeijing 100190China Hainan University Haikou 570228China the State Key Laboratory for Management and Control of Complex Systems Institute of AutomationChinese Academy of SciencesBeijing 1001901China

Powered by advanced information technology,more and more complex systems are exhibiting characteristics of the cyber-physical-social systems(CPSS).In this context,computational experiments method has emerged as a novel approach for the design,analysis,management,control,and integration of CPSS,which can realize the causal analysis of complex systems by means of“algorithmization”of“counterfactuals”.However,because CPSS involve human and social factors(e.g.,autonomy,initiative,and sociality),it is difficult for traditional design of experiment(DOE)methods to achieve the generative explanation of system *** address this challenge,this paper proposes an integrated approach to the design of computational experiments,incorporating three key modules:1)Descriptive module:Determining the influencing factors and response variables of the system by means of the modeling of an artificial society;2)Interpretative module:Selecting factorial experimental design solution to identify the relationship between influencing factors and macro phenomena;3)Predictive module:Building a meta-model that is equivalent to artificial society to explore its operating ***,a case study of crowd-sourcing platforms is presented to illustrate the application process and effectiveness of the proposed approach,which can reveal the social impact of algorithmic behavior on“rider race”.

关键词： Agent-based modeling computational experiments cyber-physical-social systems(CPSS) generative deduction generative experiments meta model

来源：评论

学校读者我要写书评

暂无评论

Multi-View Dynamic Kernelized Evidential Clustering

引用

IEEE/CAA Journal of Automatica Sinica 2024年第12期11卷 2435-2450页

作者： Jinyi Xu Zuowei Zhang Ze Lin Yixiang Chen Weiping Ding the Software Engineering Institute East China Normal UniversityShanghai 200062 Shanghai Key Laboratory of Trustworthy Computing East China Normal UniversityShanghai 200062China IEEE the School of Automation Northwestern Polytechnical UniversityXi’an 710072China the School of Artificial Intelligence and Computer Science Nantong UniversityNantong 226019 the Faculty of Data Science City University of MacaoMacao 999078China

It is challenging to cluster multi-view data in which the clusters have overlapping *** multi-view clustering methods often misclassify the indistinguishable objects in overlapping areas by forcing them into single clusters,increasing clustering *** solution,the multi-view dynamic kernelized evidential clustering method(MvDKE),addresses this by assigning these objects to meta-clusters,a union of several related singleton clusters,effectively capturing the local imprecision in overlapping *** offers two main advantages:firstly,it significantly reduces computational complexity through a dynamic framework for evidential clustering,and secondly,it adeptly handles non-spherical data using kernel techniques within its objective *** on various datasets confirm MvDKE's superior ability to accurately characterize the local imprecision in multi-view non-spherical data,achieving better efficiency and outperforming existing methods in overall performance.

关键词： Evidential clustering imprecision characterizing kernel technique multi-view clustering

来源：评论

学校读者我要写书评

暂无评论

Joint Optimization Clearing of Electro-hydrogen Coupling System for New Energy Consumption 4

Joint Optimization Clearing of Electro-hydrogen Coupling Sys...

引用

4th International Conference on Intelligent Power and Systems, ICIPS 2024

作者： Yang, Fan Wu, Hao Peng, Chen Wang, Weijie Zhang, Yajian Shanghai University School of Mechatronic Engineering and Automation China Sichuan University of Science & Engineering Artificial Intelligence Key Laboratory of Sichuan Province Yibin China Xuchang Xuji Software Technology Co. Ltd Xuchang China

ISBN: (纸本)9798331532765

In this study, a joint optimization clearing model of electro-hydrogen coupling system considering the uncertain output of new energy in electricity spot market has been proposed. Firstly, Monte-Carlo sampling and scenario reduction technologies are used to address real-time market price uncertainty. Secondly, a Cournot equilibrium model is constructed to maximize the profit of each market participant. Then, the nonlinear complementary method is used to solve the equilibrium results. Case studies have demonstrated that by optimizing the strategy of the electro-hydrogen coupling system and virtual bidding simultaneously. Meanwhile, the price gap between the day-ahead market and the real-time market can be effectively reduced, and the price of the electricity market can be more stable and reasonable. © 2024 IEEE.

关键词： Commerce

来源：评论

学校读者我要写书评

暂无评论

DA-Ada: Learning Domain-Aware Adapter for Domain Adaptive Object Detection 38

DA-Ada: Learning Domain-Aware Adapter for Domain Adaptive Ob...

引用

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Li, Haochen Zhang, Rui Yao, Hantao Zhang, Xin Hao, Yifan Song, Xinkai Li, Xiaqing Zhao, Yongwei Li, Ling Chen, Yunji Intelligent Software Research Center Institute of Software CAS Beijing China State Key Lab of Processors Institute of Computing Technology CAS Beijing China State Key Laboratory of Multimodal Artificial Intelligence Systems Institute of Automation CAS Beijing China University of Chinese Academy of Sciences Beijing China

Domain adaptive object detection (DAOD) aims to generalize detectors trained on an annotated source domain to an unlabelled target domain. As the visual-language models (VLMs) can provide essential general knowledge on unseen images, freezing the visual encoder and inserting a domain-agnostic adapter can learn domain-invariant knowledge for DAOD. However, the domain-agnostic adapter is inevitably biased to the source domain. It discards some beneficial knowledge discriminative on the unlabelled domain, ***-specific knowledge of the target domain. To solve the issue, we propose a novel Domain-Aware Adapter (DA-Ada) tailored for the DAOD task. The key point is exploiting domain-specific knowledge between the essential general knowledge and domain-invariant knowledge. DA-Ada consists of the Domain-Invariant Adapter (DIA) for learning domain-invariant knowledge and the Domain-Specific Adapter (DSA) for injecting the domain-specific knowledge from the information discarded by the visual encoder. Comprehensive experiments over multiple DAOD tasks show that DA-Ada can efficiently infer a domain-aware visual encoder for boosting domain adaptive object detection. Our code is available at https://***/Therock90421/DA-Ada. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

FedDGL: Federated Dynamic Graph Learning for Temporal Evolution and Data Heterogeneity 16

FedDGL: Federated Dynamic Graph Learning for Temporal Evolut...

引用

16th Asian Conference on Machine Learning, ACML 2024

作者： Xie, Zaipeng Li, Likun Chen, Xiangbin Yu, Hao Huang, Qian Key Laboratory of Water Big Data Technology of Ministry of Water Resources Hohai University Nanjing China College of Computer Science and Software Engineering Hohai University Nanjing China College of Artificial Intelligence and Automation Hohai University Nanjing China

Federated graph learning enhances federated learning by enabling privacy-preserving collaborative training on distributed graph data. While traditional methods are effective in managing data heterogeneity, they typically assume static graph structures, overlooking the dynamic nature of real-world graphs. Integrating federated graph learning with dynamic graph neural networks addresses this issue but often fails to retain previously acquired knowledge, limiting generalization for both global and personalized models. This paper proposes FedDGL, a novel framework designed to address temporal evolution, and data heterogeneity in federated dynamic graph learning. Unlike conventional approaches, FedDGL captures temporal dynamics through a global knowledge distillation technique and manages client heterogeneity using a global prototype-based regularization method. The framework employs contrastive learning to generate global prototypes, enhancing feature representation while utilizing a prototype similarity-based personalized aggregation strategy for effective adaptation to local and global data distributions. Experiments on multiple benchmark datasets show that FedDGL achieves significant performance improvements over state-of-the-art methods, with up to 9.02% and 8.77% gains in local and global testing, respectively, compared to FedAvg. These results highlight FedDGL’s effectiveness in improving personalized and global model performance in dynamic, heterogeneous federated graph learning scenarios. © 2024 Z. Xie, L. Li, X. Chen, H. Yu & Q. Huang.

关键词： Dynamic Graph Learning Federated Graph Learning Knowledge Distillation Personalized Aggregation Prototype Learning Temporal Evolution

来源：评论

学校读者我要写书评

暂无评论

Collaborative Decision-Making Processes Analysis of Service Ecosystem: A Case Study of Academic Ecosystem Involution 19th

Collaborative Decision-Making Processes Analysis of Service ...

引用

19th EAI International Conference on Collaborative Computing: Networking, Applications and Worksharing, CollaborateCom 2023

作者： Yan, Xiangpei Xue, Xiao Peng, Chao Liu, Donghua Feng, Zhiyong Xiao, Wang School of Computer Software College of Intelligence and Computing Tianjin University Tianjin China China Waterborne Transport Research Institute Tianjin China State Key Laboratory of Management and Control for Complex Systems Institute of Automation Chinese Academy of Sciences Beijing China

ISBN: (纸本)9783031545306

With the collaboration of several intelligent services, a crowd intelligence service network has been formed, and a service ecosystem has gradually emerged. As a novel service organization model, the Service Ecosystem (SE) can provide more sophisticated, precise, and thorough services and has attracted widespread attention. However, it also brings negative effects such as involution, and information cocoon room. Thus, how to analyze the collaborative decision-making mechanism between the SE regulation algorithm and the crowd intelligence group, exploring the reasons behind the negative effects, and finding effective intervention strategies have become problems in this field. To solve the challenges, we propose a Computational Experiments-based method Decision-making processes Analysis model in SE, namely CEDA. The proposed CEDA model consists of three modules: the autonomous evolution mechanism module, the learning evolution mechanism module, and the collaborative decision-making analysis module. Among them, the computational experiments can provide a customized test environment for the analysis of collaborative decision-making processes and find out the appropriate intervention strategy. Finally, the validity of the CEDA model is verified through the case of academic ecosystem involution. The results show that computational experiments can provide new ideas and paths for collaborative decision-making processes analysis. © ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2024.

关键词： Ecosystems

来源：评论

学校读者我要写书评

暂无评论

Data Augmentation for Vulnerability Detection Based on Code Refactoring and Mixup 35

Data Augmentation for Vulnerability Detection Based on Code ...

引用

35th IEEE International Symposium on software Reliability Engineering Workshops, ISSREW 2024

作者： Xing, Ying Huang, Jiaqi Wang, Guilong You, Yehao Yang, Bin Li, Xiaofeng Luo, Yixing Beijing University of Posts and Telecommunications School of Intelligent Engineering and Automation Beijing China Beijing University of Posts and Telecommunications School of Artifcial Intelligence Beijing China Yunnan Key Laboratory of Software Engineering Yunnan China China Unicom Research Institute Beijing China Beijing Institute of Control Engineering Software Development Center Beijing China

ISBN: (纸本)9798350367041

In recent years, the number of software vulnerabilities has gradually increased, posing a threat to the security of software, and identifying whether there are vulnerabilities in software is crucial for assuring its quality. With the development of deep learning, vulnerability detection based on deep learning has achieved good results. However, real-world software vulnerability datasets often suffer from class imbalance and a lack of labeled data, limiting the performance of deep learning models. Therefore, this paper proposes a data augmentation method to alleviate these two problems. Firstly, we treats samples with vulnerabilities in the dataset as code, and generates more samples with vulnerabilities through code refactoring to alleviate class imbalance issues. Secondly, we use the mixup method to perform linear interpolation on the embedding vectors which represent codes, and add the interpolated vectors into the training set, thereby increasing the amount of labeled training data. To evaluate our method, we conducted experiments on three projects of the Promise dataset. The experimental results show that our data augmentation method can effectively improve the effect of vulnerability detection. © 2024 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

CancerSRT:a spatially resolved transcriptomics database for human cancers

引用

Journal of Genetics and Genomics 2024年第12期51卷 1505-1508页

作者： Yuying Huo Jiakang Wang Chengcheng Liu Jinxia Wang Chen Wang Wenbo Guo Zhiyuan Yuan Tiantian Guo Jin Gu Xiangyu Li School of Software Engineering Beijing Jiaotong UniversityBeijing 100044China Department of Biological Sciences Center for Systems BiologyThe University of Texas at DallasRichardsonTX 75080USA MOE Key Laboratory of Bioinformatics BNRIST Bioinformatics DivisionDepartment of AutomationTsinghua UniversityBeijing 100084China Institute of Science and Technology for Brain-Inspired Intelligence Fudan UniversityShanghai 200433China

Human cancer is one of the leading causes of death *** heterogeneity and complex microenvironment are major challenges for anti-cancer treatment.A better understanding of the tumor heterogeneity might contribute to more precise diagnosis and *** advances in single-cell RNA-sequencing(scRNA-seq)have provided valuable insights into cell fate determination and development in cancer,but the main limitation is that cellular spatial information is ***,spatially resolved transcriptomics technologies have enabled gene expression profiling with spatial coordinates in tissues,which opens up new avenues for deciphering the cancer spatial structure and accelerating oncological research.

关键词： diagnosis spatially resolved

来源：评论

学校读者我要写书评

暂无评论

A Novel Networked Predictive Control Method for Systems with Random Communication Constraints

引用

Journal of Systems Science & Complexity 2021年第4期34卷 1364-1378页

作者： PANG Zhonghua BAI Chuandong LIU Guoping HAN Qinglong ZHANG Xianming Key Laboratory of Fieldbus Technology and Automation of Beijing North China University of TechnologyBeijing 100144China Department of Artificial Intelligence and Automation Wuhan UniversityWuhan 430072China School of Software and Electrical Engineering Swinburne University of TechnologyMelbourneVIC 3122Australia

This paper presents a novel observer-based predictive control method for networked systems where random network-induced delays,packet disorders and packet dropouts in both feedback and forward channels are *** proposed method has three significant features:i)A concept of destination-based lumped(DBL)delay is introduced to represent the combined effects of random communication constraints in each channel;ii)in view of different natures of the random DBL delays in the feedback and forward channels,different compensation schemes are designed;and iii)it is actual control inputs rather than predicted ones that are employed to generate future control signals based on the latest system state estimate available in the *** the resulting closed-loop system,a necessary and sufficient stability condition is derived,which is less conservative and also independent of random communication constraints in both *** results are provided to demonstrate the effectiveness of the proposed method.

关键词： Networked control systems(NCSs) network-induced delay networked predictive control packet disorder packet dropout

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：