检索结果-内蒙古大学图书馆

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Guo, Pengxin Zeng, Shuang Chen, Wenhao Zhang, Xiaodan Ren, Weihong Zhou, Yuyin Qu, Liangqiong School of Computing and Data Science The University of Hong Kong Hong Kong Department of Mathematics The University of Hong Kong Hong Kong College of Computer Science Beijing University of Technology China School of Mechanical Engineering and Automation Harbin Institute of Technology Shenzhen China Department of Computer Science and Engineering UC Santa Cruz United States

ISBN: (纸本)157735897X

Federated Learning (FL) aims to protect data privacy by enabling clients to collectively train machine learning models without sharing their raw data. However, recent studies demonstrate that information exchanged during FL is subject to Gradient Inversion Attacks (GIA) and, consequently, a variety of privacy-preserving methods have been integrated into FL to thwart such attacks, such as Secure Multi-party computing (SMC), Homomorphic Encryption (HE), and Differential Privacy (DP). Despite their ability to protect data privacy, these approaches inherently involve substantial privacy-utility trade-offs. By revisiting the key to privacy exposure in FL under GIA, which lies in the frequent sharing of model gradients that contain private data, we take a new perspective by designing a novel privacy preserve FL framework that effectively “breaks the direct connection” between the shared parameters and the local private data to defend against GIA. Specifically, we propose a Hypernetwork Federated Learning (HyperFL) framework that utilizes hypernetworks to generate the parameters of the local model and only the hypernetwork parameters are uploaded to the server for aggregation. Theoretical analyses demonstrate the convergence rate of the proposed HyperFL, while extensive experimental results show the privacy-preserving capability and comparable performance of HyperFL. Copyright © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

SoftGPT: Learn Goal-oriented Soft Object Manipulation Skills by Generative Pre-trained Heterogeneous Graph Transformer

arXiv

引用

arXiv 2023年

作者： Liu, Junjia Li, Zhihao Lin, Wanyu Calinon, Sylvain Tan, Kay Chen Chen, Fei The Department of Mechanical and Automation Engineering T-Stone Robotics Institute The Chinese University of Hong Kong Hong Kong The Department of Computing The Hong Kong Polytechnic University Hong Kong The Idiap Research Institute Martigny Switzerland

Soft object manipulation tasks in domestic scenes pose a significant challenge for existing robotic skill learning techniques due to their complex dynamics and variable shape characteristics. Since learning new manipulation skills from human demonstration is an effective way for robot applications, developing prior knowledge of the representation and dynamics of soft objects is necessary. In this regard, we propose a pre-trained soft object manipulation skill learning model, namely SoftGPT, that is trained using large amounts of exploration data, consisting of a three-dimensional heterogeneous graph representation and a GPT-based dynamics model. For each downstream task, a goal-oriented policy agent is trained to predict the subsequent actions, and SoftGPT generates the consequences of these actions. Integrating these two approaches establishes a thinking process in the robot’s mind that provides rollout for facilitating policy learning. Our results demonstrate that leveraging prior knowledge through this thinking process can efficiently learn various soft object manipulation skills, with the potential for direct learning from human demonstrations. Copyright © 2023, The Authors. All rights reserved.

关键词： Robot applications

来源：评论

学校读者我要写书评

暂无评论

Total Variation Constrained Non-Negative Matrix Factorization for Medical Image Registration

引用

IEEE/CAA Journal of Automatica Sinica 2021年第5期8卷 1025-1037页

作者： Chengcai Leng Hai Zhang Guorong Cai Zhen Chen Anup Basu the School of Mathematics Northwest UniversityXi’an 710127 the Institute of Automation Chinese Academy of SciencesBeijing 100190China the Department of Computing Science University of AlbertaEdmontonAB T6G 2E8Canada the College of Computer Engineering Jimei UniversityXiamen 361021China the School of Measuring and Optical Engineering Nanchang Hangkong UniversityNanchang 330063China

This paper presents a novel medical image registration algorithm named total variation constrained graphregularization for non-negative matrix factorization(TV-GNMF).The method utilizes non-negative matrix factorization by total variation constraint and graph *** main contributions of our work are the ***,total variation is incorporated into NMF to control the diffusion *** purpose is to denoise in smooth regions and preserve features or details of the data in edge regions by using a diffusion coefficient based on gradient ***,we add graph regularization into NMF to reveal intrinsic geometry and structure information of features to enhance the discrimination ***,the multiplicative update rules and proof of convergence of the TV-GNMF algorithm are *** conducted on datasets show that the proposed TV-GNMF method outperforms other state-of-the-art algorithms.

关键词： Data clustering dimension reduction image registration non-negative matrix factorization(NMF) total variation(TV)

来源：评论

学校读者我要写书评

暂无评论

Approaching Prosumer Social Optimum via Energy Sharing With Proof of Convergence

Approaching Prosumer Social Optimum via Energy Sharing With ...

引用

IEEE General Meeting Power& Energy Society

作者： Yue Chen Changhong Zhao Steven Low Shengwei Mei Department of Mechanical and Automation Engineering The Chinese University of Hong Kong Department of Electrical Engineering Tsinghua University Department of Information Engineering The Chinese University of Hong Kong Department of Computing and Mathematical Sciences California Institute of Technology

ISBN: (数字)9781665408233

ISBN: (纸本)9781665408240

With the advent of prosumers, the traditional centralized operation may become impracticable due to computational burden, privacy concerns, and conflicting interests. In this article, an energy sharing mechanism is proposed to accommodate prosumers' strategic decision-making on their self-production and demand in the presence of capacity constraints. Under this setting, prosumers play a generalized Nash game. We prove main properties of the game: an equilibrium exists and is partially unique; no prosumer is worse off by energy sharing and the price-of-anarchy is O(1/I) where I is the number of prosumers. In particular, the PoA tends to 1 with a growing number of prosumers, meaning that the resulting total cost under the proposed energy sharing approaches social optimum. We prove that the corresponding prosumers' strategies converge to the social optimal solution as well. Finally we propose a bidding process and prove that it converges to the energy sharing equilibrium under mild conditions. Illustrative examples are provided to validate the results.

关键词： Privacy Costs Decision making Games Convergence

来源：评论

学校读者我要写书评

暂无评论

MFANet: Mixed Feature Attention Network for Port Lane Detection

MFANet: Mixed Feature Attention Network for Port Lane Detect...

引用

Chinese Control Conference (CCC)

作者： Jinwei Zhang Jinya Su Dewei Yi Jun Yang Key Laboratory of Measurement and Control of Complex Systems of Engineering Ministry of Education and the School of Automation Southeast University Nanjing China Department of Computing Science University of Aberdeen Aberdeen U.K.

With the advancement of industrial automation and artificial intelligence technology, unmanned port autonomy has gained increasing attention. Port automatic driving technology is a critical component of unmanned autonomous systems, enabling rubber tire gantry cranes (RTGs) to autonomously navigate by detecting port lanes and implementing control commands. In this study, a deep learning-based method is proposed for detecting port lanes in complex port scenes. The method employs the Inception module and attention mechanism to enhance lane feature extraction, and a structural loss is introduced to explicitly constrain the detection results and incorporate the prior characteristics of port lanes. A novel Mixed Feature Attention Network (MFANet) is proposed to implement the method for port lane detection. Experimental results demonstrate that MFANet effectively improves the accuracy of port lane detection while maintaining high computational efficiency. Furthermore, the performance of MFANet is evaluated in a real-world application scenario.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Empirical Privacy Variance

arXiv

引用

arXiv 2025年

作者： Hu, Yuzheng Wu, Fan Xian, Ruicheng Liu, Yuhang Zakynthinou, Lydia Kamath, Pritish Zhang, Chiyuan Forsyth, David Siebel School of Computing and Data Science University of Illinois Urbana-Champaign United States Institute of Automation Chinese Academy of Sciences China Department of Electrical Engineering and Computer Science University of California Berkeley United States Google Research United States

We propose the notion of empirical privacy variance and study it in the context of differentially private fine-tuning of language models. Specifically, we show that models calibrated to the same (Ε, δ)-DP guarantee using DP-SGD with different hyperparameter configurations can exhibit significant variations in empirical privacy, which we quantify through the lens of memorization. We investigate the generality of this phenomenon across multiple dimensions and discuss why it is surprising and relevant. Through regression analysis, we examine how individual and composite hyperparameters influence empirical privacy. The results reveal a no-free-lunch trade-off: existing practices of hyperparameter tuning in DP-SGD, which focus on optimizing utility under a fixed privacy budget, often come at the expense of empirical privacy. To address this, we propose refined heuristics for hyperparameter selection that explicitly account for empirical privacy, showing that they are both precise and practically useful. Finally, we take preliminary steps to understand empirical privacy variance. We propose two hypotheses, identify limitations in existing techniques like privacy auditing, and outline open questions for future research. Copyright © 2025, The Authors. All rights reserved.

关键词： Differential privacy

来源：评论

学校读者我要写书评

暂无评论

Research on perimeter intrusion monitoring technology of high-speed railway based on multi-source heterogeneous data fusion 24

Research on perimeter intrusion monitoring technology of hig...

引用

Proceedings of the 2024 4th International Conference on Big Data, Artificial Intelligence and Risk Management

作者： Wentao Li Tao Yu Bo Zhang Huijin Fu Graduate Department China Academy of Railway Sciences Beijing China School of Automation and Software Engineering Shanxi University Taiyuan Shanxi China Institute of Computing Technologies China Academy of Railway Sciences Group Co. Ltd. Beijing China

ISBN: (纸本)9798400709753

Aiming at the defects of a single protection technology for high-speed railway perimeter intrusion, such as poor environmental adaptability and difficult technical shortcomings to improve, this paper conducted an in-depth study on railway perimeter intrusion monitoring technology based on multi-source heterogeneous data fusion, and designed a monitoring technology scheme of system coupling and collaborative linkage integration. In order to solve the technical shortcomings of a single sensor, the whole scene high-precision monitoring under bad weather is realized. Due to their geographical characteristics, the tunnel entrance and the road bridge over the railway are prone to foreign body intrusion caused by geological disasters or human factors, so they are the key protection areas for the perimeter intrusion of high-speed railway. In this paper, a high-speed rail perimeter intrusion monitoring technology based on multi-source heterogeneous data fusion is studied for two typical running scenarios of tunnel entrance and upper span railway bridge. The test results show that the proposed technology can meet the demand of accurate and timely railway perimeter intrusion alarm.

关键词： Data fusion High-speed railway Multi-source heterogeneous Object detection Perimeter intrusion

来源：评论

学校读者我要写书评

暂无评论

State-Conditioned Adversarial Subgoal Generation

arXiv

引用

arXiv 2022年

作者： Wang, Vivienne Huiling Pajarinen, Joni Wang, Tinghuai Kämäräinen, Joni-Kristian Computing Sciences Tampere University Finland Department of Electrical Engineering and Automation Aalto University Finland Huawei Helsinki Research Center Finland

Hierarchical reinforcement learning (HRL) proposes to solve difficult tasks by performing decision-making and control at successively higher levels of temporal abstraction. However, off-policy HRL often suffers from the problem of a non-stationary high-level policy since the low-level policy is constantly changing. In this paper, we propose a novel HRL approach for mitigating the non-stationarity by adversarially enforcing the high-level policy to generate subgoals compatible with the current instantiation of the low-level policy. In practice, the adversarial learning is implemented by training a simple state conditioned discriminator network concurrently with the high-level policy which determines the compatibility level of subgoals. Comparison to state-of-the-art algorithms shows that our approach improves both learning efficiency and performance in challenging continuous control tasks. © 2022, CC BY.

关键词： Decision making

来源：评论

学校读者我要写书评

暂无评论

Three-View Cotton Flower Counting through Multi-Object Tracking and Multi-Modal Imaging

Three-View Cotton Flower Counting through Multi-Object Track...

引用

2023 American Society of Agricultural and Biological Engineers Annual International Meeting, ASABE 2023

作者： Tan, Chenjiao Li, Changying Sun, Jin Song, Huaibo College of Mechanical and Electronic Engineering Northwest A&F University Yangling China Key Laboratory of Agricultural Internet of Things Ministry of Agriculture and Rural Affairs Yangling China Shaanxi Key Laboratory of Agricultural Information Perception and Intelligent Service Yangling China Bio-sensing Automation and Intelligence Lab Department of Agricultural and Biological Engineering University of Florida GainesvilleFL United States School of Computing University of Georgia AthensGA United States

ISBN: (纸本)9781713885887

Monitoring the number of cotton flowers can provide important information for breeders to assess the flowering time and the productivity of genotypes because flowering marks the transition from vegetative growth to reproductive growth and impacts the number of bolls developed. Typically, manual counting is difficult, time-consuming, and tedious. To count cotton flowers efficiently and accurately, we propose a multi-view and multi-modal imaging approach based on object tracking using both RGB and depth images collected by three cameras mounted on a ground robotic platform. Tracking by detection algorithm was employed to track flowers from three views at the same time. Furthermore, a YOLOv8 object detector was trained to detect flowers in images and a deep learning-based optical flow model RAFT (Recurrent All-pairs Field Transforms) was used to estimate motion between two adjacent frames. IoU and distance costs were used to associate flowers in the tracking algorithm, and tracked flowers were segmented to obtain pixel positions for depth extraction. These positions from two side views were then projected onto the middle image coordinate using camera calibration parameters. Finally, all flowers in the middle image coordinate were clustered by a constrained hierarchy clustering algorithm to remove duplicated counts. The results showed that the mAP of trained YOLOv8x was 96.4%. The counting results of the developed method were highly correlated with those of manual counts with an R2 of 0.92. Besides, the Mean Relative Error of all testing videos was 6.22%. Overall, the proposed method provides an efficient and effective approach to solving the cotton flower counting from multiple views, which is beneficial for breeders to dissect genetic mechanisms of flowering time with an unprecedented spatial and temporal resolution. © 2023 ASABE Annual International Meeting. All Rights Reserved.

关键词： Optical flows

来源：评论

学校读者我要写书评

暂无评论

A Review and Benchmark on State-of-the-Art Steel Defects Detection

引用

SN Computer Science 2024年第1期5卷 114页

作者： Chazhoor, Anthony Ashwin Peter Ho, Edmond S. L. Gao, Bin Woo, Wai Lok Computer and Information Science Department Northumbria University ELB 108 Ellison Pl Newcastle upon Tyne NE1 8ST United Kingdom School of Computing Science University of Glasgow 18 Lilybank Gardens Glasgow G12 8RZ United Kingdom School of Automation Engineering University of Electronic Science and Technology of China 4 1st Ring Rd East 2 Section Chenghua District Sichuan Chengdu 610056 China

Steel, a critical material in construction, automobile, and railroad manufacturing industries, often presents defects that can lead to equipment failure, significant safety risks, and costly downtime. This research aims to evaluate the performance of state-of-the-art object detection models in detecting defects on steel surfaces, a critical task in industries such as railroad and automobile manufacturing. The study addresses the challenges of limited defect data and lengthy model training times. Five existing state-of-the-art object detection models (faster R-CNN, deformable DETR, double head R-CNN, Retinanet, and deformable convolutional network) were benchmarked on the Northeastern University (NEU) steel dataset. The selection of models covers a broad spectrum of methodologies, including two-stage detectors, single-stage detectors, transformers, and a model incorporating deformable convolutions. The deformable convolutional network achieved the highest accuracy of 77.28% on the NEU dataset following a fivefold cross-validation method. Other models also demonstrated notable performance, with accuracies within the 70–75% range. Certain models exhibited particular strengths in detecting specific defects, indicating potential areas for future research and model improvement. The findings provide a comprehensive foundation for future research in steel defect detection and have significant implications for practical applications. The research could improve quality control processes in the steel industry by automating the defect detection task, leading to safer and more reliable steel products and protecting workers by removing the human factor from hazardous environments. © 2023, The Author(s).

关键词： Automated defect detection Deep learning models Steel defect detection Steel surface

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：