检索结果-内蒙古大学图书馆

IEEE Annual Southern Power Electronics Conference (SPEC)

作者： Sarat Singamneni Rodney Badcock Grant Lumsden Jeong Sengkwon Alan Caughley Duleepa Thrimawithana Stijn Tissink Malaya Prasad Behera Yifan Lv Aaron Wadsworth Eloise Cameron-Smith Vikash Kumar Department of Mechanical Engineering Auckland University of Technology New Zealand Paihau-Robinson Research Institute Victoria University of Wellington Wellington New Zealand Cryogenics Engineering Laboratory Korean Advanced Institute of Technology Seoul Korea Advanced Engineering and Materials Callaghan Innovation Christchurch New Zealand Dept Elect. Comp. and Software Eng. University of Auckland Auckland New Zealand Department of Mechanical Engineering Auckland University of Technology Auckland New Zealand Dept Elect. Comp. and Software Eng University of Auckland Auckland New Zealand

ISBN: (数字)9798350351156

ISBN: (纸本)9798350351163

GaN HEMTs on circuit boards used for power electronics applications suffer from the generation of excessive heat that may eventually destroy them. Both immersion and conduction (using a cold plate) methods for cooling have been evaluated in the past. The objective of this work is to design and build a cold plate solution to this thermal problem by means of forced convection through optimised cooling channels integrated into a cold plate. Considering the comp.ex and convoluted shapes these fluid-flow channels take within enclosures that rest on the heat dissipating surface of the GaN chip, additive manufacturing based on powder bed fusion processing of a thermally conductive metal is a part of the solution. Finite volume (FV) simulations based on ANSYS Fluent module are used for evaluating thermal fields within the heat exchanger block resting on the GaN HEMTs, taken as the plate heat source. Initial impressions from the FV results indicate the design alterations evaluated to be suitable for better thermal management of GaN HEMTs. Physical prototyping of the cooling block was also successful based on additive consolidation of metallic powders by selective laser melting.

关键词： Thermal management of electronics Heating systems Powders Thermal resistance Cold plates HEMTs Thermal conductivity Three-dimensional printing Power electronics MODFETs

来源：评论

学校读者我要写书评

暂无评论

CODIS: Benchmarking Context-Dependent Visual comp.ehension for Multimodal Large Language Models

arXiv

引用

arXiv 2024年

作者： Luo, Fuwen Chen, Chi Wan, Zihao Kang, Zhaolu Yan, Qidong Li, Yingjie Wang, Xiaolong Wang, Siyu Wang, Ziyue Mi, Xiaoyue Li, Peng Ma, Ning Sun, Maosong Liu, Yang Dept. of Comp. Sci. & Tech. Institute for AI Tsinghua University Beijing China Tsinghua University Beijing China Shanghai Artificial Intelligence Laboratory Shanghai China Jiangsu Collaborative Innovation Center for Language Competence Jiangsu China Key Laboratory of Linguistic and Cultural Computing Ministry of Education Northwest Minzu University China College of Software Jilin University China Institute of Computing Technology Chinese Academy of Sciences China

Multimodal large language models (MLLMs) have demonstrated promising results in a variety of tasks that combine vision and language. As these models become more integral to research and applications, conducting comp.ehensive evaluations of their capabilities has grown increasingly important. However, most existing benchmarks fail to consider that, in certain situations, images need to be interpreted within a broader context. In this work, we introduce a new benchmark, named as CODIS, designed to assess the ability of models to use context provided in free-form text to enhance visual comp.ehension. Our findings indicate that MLLMs consistently fall short of human performance on this benchmark. Further analysis confirms that these models struggle to effectively extract and utilize contextual information to improve their understanding of images. This underscores the pressing need to enhance the ability of MLLMs to comp.ehend visuals in a context-dependent manner. View our project website at https://***/CODIS. © 2024, CC BY.

关键词： Benchmarking

来源：评论

学校读者我要写书评

暂无评论

Text-to-Image Diffusion Models can be Easily Backdoored through Multimodal Data Poisoning

arXiv

引用

arXiv 2023年

作者： Zhai, Shengfang Dong, Yinpeng Shen, Qingni Pu, Shi Fang, Yuejian Su, Hang School of Software and Microelectronics Peking University China Department of Computer Science & Technology Tsinghua University China ShengShu Beijing China School of Software and Microelectronics National Engineering Research Center for Software Engineering PKU-OCTA Laboratory for Blockchain and Privacy Computing Peking University Beijing100871 China Dept. of Comp. Sci. and Tech. Institute for AI Tsinghua-Bosch Joint ML Center THBI Lab BNRist Center Tsinghua University Beijing100084 China

With the help of conditioning mechanisms, the state-of-the-art diffusion models have achieved tremendous success in guided image generation, particularly in text-to-image synthesis. To gain a better understanding of the training process and potential risks of text-to-image synthesis, we perform a systematic investigation of backdoor attack on text-to-image diffusion models and propose BadT2I, a general multimodal backdoor attack framework that tampers with image synthesis in diverse semantic levels. Specifically, we perform backdoor attacks on three levels of the vision semantics: Pixel-Backdoor, Object-Backdoor and Style-Backdoor. By utilizing a regularization loss, our methods efficiently inject backdoors into a large-scale text-to-image diffusion model while preserving its utility with benign inputs. We conduct empirical experiments on Stable Diffusion, the widely-used text-to-image diffusion model, demonstrating that the large-scale diffusion model can be easily backdoored within a few fine-tuning steps. We conduct additional experiments to explore the impact of different types of textual triggers, as well as the backdoor persistence during further training, providing insights for the development of backdoor defense methods. Besides, our investigation may contribute to the copyright protection of text-to-image models in the future. Our Code: https://***/sf-zhai/BadT2I. Copyright © 2023, The Authors. All rights reserved.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Unilaterally Aggregated Contrastive Learning with Hierarchical Augmentation for Anomaly Detection

Unilaterally Aggregated Contrastive Learning with Hierarchic...

引用

International Conference on comp.ter Vision (ICCV)

作者： Guodong Wang Yunhong Wang Jie Qin Dongming Zhang Xiuguo Bao Di Huang State Key Laboratory of Software Development Environment Beihang University Beijing China School of Computer Science and Engineering Beihang University Beijing China College of Computer Science and Technology NUAA Nanjing China Natl. Comp. Net. Emer. Resp. Tech. Team/Coord. Ctr. of China Beijing China

Anomaly detection (AD), aiming to find samples that deviate from the training distribution, is essential in safety-critical applications. Though recent self-supervised learning based attempts achieve promising results by creating virtual outliers, their training objectives are less faithful to AD which requires a concentrated inlier distribution as well as a dispersive outlier distribution. In this paper, we propose Unilaterally Aggregated Contrastive Learning with Hierarchical Augmentation (UniCon-HA), taking into account both the requirements above. Specifically, we explicitly encourage the concentration of inliers and the dispersion of virtual outliers via supervised and unsupervised contrastive losses, respectively. Considering that standard contrastive data augmentation for generating positive views may induce outliers, we additionally introduce a soft mechanism to re-weight each augmented inlier according to its deviation from the inlier distribution, to ensure a purified concentration. Moreover, to prompt a higher concentration, inspired by curriculum learning, we adopt an easy-to-hard hierarchical augmentation strategy and perform contrastive aggregation at different depths of the network based on the strengths of data augmentation. Our method is evaluated under three AD settings including unlabeled one-class, unlabeled multi-class, and labeled multi-class, demonstrating its consistent superiority over other comp.titors.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Unilaterally Aggregated Contrastive Learning with Hierarchical Augmentation for Anomaly Detection

arXiv

引用

arXiv 2023年

作者： Wang, Guodong Wang, Yunhong Qin, Jie Zhang, Dongming Bao, Xiuguo Huang, Di State Key Laboratory of Software Development Environment Beihang University Beijing China School of Computer Science and Engineering Beihang University Beijing China College of Computer Science and Technology NUAA Nanjing China Natl. Comp. Net. Emer. Resp. Tech. Team/Coord. Ctr. of China Beijing China

关键词： Anomaly detection

来源：评论

学校读者我要写书评

暂无评论

Reliability Analysis of Energy-Efficient Parallel Storage Systems

Reliability Analysis of Energy-Efficient Parallel Storage Sy...

引用

IEEE International Conference on High Performance comp.ting and Communications (HPCC)

作者： Shu Yin Kenli Li Xiao Qin Xiaojun Ruan Haiquan Chen Jiong Xie Xiaomin Zhu School of Info. Sci. and Eng. Hunan University Changsha China Dept. of Comp. Sci. and Software Eng. Auburn University Auburn AL Dept. of Computer Science West Chester University of Pennsylvania West Chester PA Dept of Math. and Computer Sci. Valdosta State University Valdosta GA Info. and Comm. Center Inner Mongolia Electric Power Hohhot China Sci. and Tech. on Info. Systems Eng. Laboratory National University of Defense Technology Changsha China

Existing reliability models evaluate the lifetime of storage systems well. However, few models aim to analysis parallel storage systems, especially systems equipped with energy-efficient solutions. MREED model provides an inspiring reliability analysis idea for energy-efficient parallel storage systems, especially RAIDs. Unfortunately, the MREED has its disadvantage that it can evaluate RAID-0 only, in which a basic method of data-stripping is applied. In this paper, we extend the MREED model to cover the most energy-efficient RAID setups including RAID-1, which introduces disk-mirroring techniques, and RAID-5, which applies a data-parity techniques based on data-stripping schemes for a data recovery purpose. The MREED is then applied to evaluate an existing well-known power-aware RAID system-PARAID. Experimental results indicate the data locality affection and data parity impaction on the reliability of PARAID. Furthermore, we investigate a fundamental trade-off between reliability and energy-efficiency in terms of energy-saving RAID systems. A tradeoff curve is provided to justify whether it worth to sacrifice reliability for better energy efficiency.

关键词： Reliability Gears Temperature Arrays Energy storage Spinning Educational institutions

来源：评论

学校读者我要写书评

暂无评论

Practical privacy for value-added applications in vehicular ad hoc networks 5th

引用

5th International Conference on Internet and Distributed comp.ting Systems, IDCS 2012

作者： Zhang, Lei Wu, Qianhong Qin, Bo Domingo-Ferrer, Josep Shanghai Key Laboratory of Trustworthy Computing Software Engineering Institute East China Normal University Shanghai China UNESCO Chair in Data Privacy Dept. of Comp. Eng. and Maths Universitat Rovira i Virgili TarragonaCatalonia Spain Key Lab. of Aerospace Information Security and Trusted Computing Ministry of Education Wuhan University School of Computer China Department of Maths School of Science Xi’an University of Technology China

ISBN: (纸本)9783642348822

Advances in mobile networking and information processing technologies have triggered vehicular ad hoc networks (VANETs) for traffic safety and value-added applications. Most efforts have been made to address the security concerns while little work has been done to investigate security and privacy for value-added applications in VANETs. To fill this gap, we propose a value-added application, specifically, a security and privacy preserving location-based service (LBS) scheme for VANETs. For each LBS transaction, the scheme provides authentication, integrity and non-repudiation for both the service provider and the user. A user can obtain the service in an anonymous way and hence user privacy is well protected. However, a tracing procedure can be invoked to find malicious users, thereby efficiently preventing users from abusing the anonymity provided by the system. © Springer-Verlag Berlin Heidelberg 2012.

关键词： Location based services

来源：评论

学校读者我要写书评

暂无评论

Multi-institutional, multi-national studies in CSEd research: Some design considerations and trade-offs 05

Multi-institutional, multi-national studies in CSEd research...

引用

1st International comp.ting Education Research Workshop, ICER 2005

作者： Fincher, Sally Lister, Raymond Clear, Tony Robins, Anthony Tenenberg, Josh Petre, Marian Computing Laboratory University of Kent Canterbury CT2 7NF United Kingdom Faculty of Information Technology University of Technology Sydney Broadway NSW 2007 Australia School of Comp. and Inf. Sciences Auckland University of Technology Private Bag 92006 Auckland 1020 New Zealand Computer Science Department University of Otago PO Box 56 Dunedin 9015 New Zealand University of Washington Tacoma Computing and Software Systems 1900 Commerce St. Tacoma WA 98402-3100 United States Faculty of Mathematics and Computing Open University Milton Keynes MK7 6AA United Kingdom

ISBN: (纸本)1595930434

One indication of the maturation of comp.ter Science Education as a research-based discipline is the recent emergence of several large-scale studies spanning multiple institutions. This paper examines a "family" of these multi-institutional, multi-national studies, detailing core elements and points of difference in both study design and the organization of the research team, and highlighting the costs and benefits associated with the different approaches. Copyright 2005 ACM.

关键词： Engineering research

来源：评论

学校读者我要写书评

暂无评论

A dynamic model of multi-agent system

A dynamic model of multi-agent system

引用

Proceedings of 2004 International Conference on Machine Learning and Cybernetics

作者： Hu, Shan-Li Shi, Chun-Yi Dept. of Comp. Sci. and Technology Fuzhou University Fuzhou 350002 China Laboratory of Computer Science Institute of Software Chinese Academy of Sciences Beijing 100080 China Dept. of Comp. Sci. and Technology Tsinghua University Beijing 100084 China

ISBN: (纸本)0780384032

There are two purposes in this paper. First, it presents an approach to designing an agent For adapting to the cooperative requests of the agents in a multi-agent system, we distinguish between the two aspects of underlying attributes of an agent: individual and organizational. And introduce a concept "pairs of roles". The other purpose is to show how to express the behavioral semantics of an agent and a multi-agent system with the "state transition graph". The individual attributes of an agent consist of beliefs and intentions, and for intentions we adopt the "twin-subset semantics" developed by us. While the organizational attributes of an agent consist of "pairs of roles". These comp.se the state of individual agent and the relations between agents. The state transitions are produced by the behaviors of agents and materialize the dynamic characters of agent and the turn of a multi-agent system. And the behaviors arise from the mental states of agents and the situations.

关键词： Multi agent systems

来源：评论

学校读者我要写书评

暂无评论

Service selection in dynamic demand-driven Web services

Service selection in dynamic demand-driven Web services

引用

Proceedings - IEEE International Conference on Web Services

作者： Wang, Xiaoling Yue, Kun Huang, Joshua Zhexue Zhou, Aoying Dept. of Comp. Sci. and Engineering Fudan University ShangHai 200433 China E-Business Technology Institute University of Hong Kong Hong Kong Hong Kong Laboratory of Computer Science Institute of Software Chinese Academy of Science China

ISBN: (纸本)0769521673

Recently, Web services have become a new technology trend for Enterprise Application Integration (EAI) and more and more applications based on Web services are emerging. One of the problems in using Web services in business applications such as logistics is services comp.sition automatically and efficiently. In this paper, we present a Dynamic, Demand-Driven Web services Engine called D3D-Serv to implement comp.site service functionality that is used to dynamically build comp.site services from existing services according to different business logics and requirements. In this D3D_Serv framework, the most challenging function to implement is dynamic selection of service providers at run time. The highly dynamic and distributed nature of Web services often makes some service providers overloaded at certain times while others idle. To solve this problem, we propose an efficient services selection and execution strategy that is based on the queuing theory and can provide guarantees for the QOS (Quality of Service) under provider's limited resources. Preliminary experimental results have shown that this algorithm is effective.

关键词： World Wide Web

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：