检索结果-内蒙古大学图书馆

arXiv 2020年

作者： Wang, Mengzhu Zhang, Xiang Lan, Long Wang, Wei Tan, Huibin Luo, Zhigang Science and Technology on Parallel and Distributed Laboratory College of Computer National University of Defense Technology Changsha China Institute for Quantum State Key Laboratory of High Performance Computing National University of Defense Technology Changsha China DUT-RU International School of Information Science & Engineering Dalian University of Technology DalianLiaoning116000 China Department of Science and Technology on Parallel and Distributed Processing National University of Defense Technology Changsha China

Reducing feature redundancy has shown beneficial effects for improving the accuracy of deep learning models, thus it is also indispensable for the models of unsupervised domain adaptation (UDA). Nevertheless, most recent efforts in the field of UDA ignore this point. Moreover, main schemes realizing this in general independent of UDA purely involve a single domain, thus might not be effective for cross-domain tasks. In this paper, we emphasize the significance of reducing feature redundancy for improving UDA in a bi-level way. For the first level, we try to ensure compact domain-specific features with a transferable decorrelated normalization module, which preserves specific domain information whilst easing the side effect of feature redundancy on the sequel domain-invariance. In the second level, domain-invariant feature redundancy caused by domain-shared representation is further mitigated via an alternative brand orthogonality for better generalization. These two novel aspects can be easily plugged into any BN-based backbone neural networks. Specifically, simply applying them to ResNet50 has achieved competitive performance to the state-of-the-arts on five popular benchmarks. Our code will be available at https://***/dreamkily/gUDA. © 2020, CC BY.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

RuleRAG: Rule-Guided Retrieval-Augmented Generation with Language Models for Question Answering

arXiv

引用

arXiv 2024年

作者： Chen, Zhongwu Xu, Chengjin Wang, Dingmin Huang, Zhen Dou, Yong Jiang, Xuhui Guo, Jian National Key Laboratory of Parallel and Distributed Computing College of Computer Science and Technology National University of Defense Technology China IDEA Research International Digital Economy Academy Department of Computer Science University of Oxford United Kingdom

Retrieval-augmented generation (RAG) has shown promising potential in knowledge intensive question answering (QA). However, existing approaches only consider the query itself, neither specifying the retrieval preferences for the retrievers nor informing the generators of how to refer to the retrieved documents for the answers, which poses a significant challenge to the QA performance. To address these issues, we propose Rule-guided Retrieval-Augmented Generation with LMs, which explicitly introduces rules for in-context learning (RuleRAG-ICL) to guide retrievers to recall related documents in the directions of rules and uniformly guide generators to reason attributed by the same rules. The combination of queries and rules can be used as fine-tuning data to update retrievers and generators, achieving better rule-based instruction-following ability (RuleRAG-FT). Moreover, most existing RAG datasets were constructed without considering rules and Knowledge Graphs (KGs) are recognized as providing high-quality rules. Therefore, we construct five rule-aware RAG benchmarks for QA, RuleQA, based on KGs to stress the significance of retrieval and reasoning with rules. Experiments on RuleQA demonstrate RuleRAG-ICL improves the retrieval quality of +89.2% in Recall@10 and answer accuracy of +103.1% in Exact Match, and RuleRAG-FT yields more enhancement. In addition, experiments on four existing RAG datasets show RuleRAG is also effective by offering rules in RuleQA to them, further proving the generalization of rule guidance in RuleRAG. Code and RuleQA are at https://***/r/RuleRAG. Copyright © 2024, The Authors. All rights reserved.

关键词： Question answering

来源：评论

学校读者我要写书评

暂无评论

An Efficient Label Routing on High-Radix Interconnection Networks

An Efficient Label Routing on High-Radix Interconnection Net...

引用

International Conference on parallel and distributed Systems (ICPADS)

作者： Fei Lei Dezun Dong Xiangke Liao College of Computer National University of Defense Technology China National Laboratory for Parallel and Distributed Processing National University of Defense Technology China Collaborative Innovation Center of High Performance Computing National University of Defense Technology China

ISBN: (纸本)9781538632086

Cost-effective adaptive routing has a significant impact on overall performance for high-radix hierarchical topologies, such as Dragonfly, which achieve a lower network diameter than traditional topologies, Torus and Fat tree, but exhibit a lower degree of adaptiveness for shortest-path rout- ing. Existing adaptive routing methods for those hierarchical topologies improve the adaptiveness by increasing path length, i.e. local or global adaptive routing, and thus suffer from complex and costly deadlock avoidance. This work aims to maximize the routing adaptiveness at the minimum cost of deadlock avoidance. We propose a label routing method for high-radix hierarchical networks. This label routing utilizes a co-design methodology and coordinates the two pipelines, input queue and routing computation, in the router microarchitec- ture. Packets in the input buffer are labeled by our routing algorithm depending on network states. We reorganize the input buffer and develop a label routing algorithm, named Green-Red Routing, GRR. GRR relaxes the requirement of using virtual channels to eliminate routing deadlock, and mitigates buffer resources dedicated to deadlock avoidance. GRR manages the buffer resources and balance its utilization elaborately, and achieve fully adaptive routing efficiently. We conduct extensive experiments to evaluate the performance of GRR on Dragonfly and compare it with state-of-the-art works. The results show that GRR achieves 10%-35% higher performance than existing routing algorithms under most traffic patterns.

关键词： Routing System recovery Network topology Topology Resource management Adaptive systems Organizations

来源：评论

学校读者我要写书评

暂无评论

FedHybrid: Hierarchical Hybrid Training for High-Performance Federated Learning

FedHybrid: Hierarchical Hybrid Training for High-Performance...

引用

IEEE Communications Society Conference on Sensor and Ad Hoc Communications and Networks (SECON)

作者： Tao Chang Li Li Meihan Wu Wei Yu Xiaodong Wang National Key Laboratory of Parallel and Distributed Computing National University of Defense Technology China State Key Laboratory of Internet of Things for Smart City University of Macau China The 30th Research Institute of China Electronics Technology Group Corporation China

Federated Learning coordinates multiple devices to train a shared model while preserving data privacy. Despite its potential benefit, the increasing number of participating devices poses new challenges to the deployment in real-world cases. The highly limited amount of data located on each device coupled with significantly unbalanced data across different devices severely impede the performance of the shared model and the overall training progress at the same *** this paper, we propose FedHybrid, a hierarchical hybrid training framework for high-performance Federated Learning on a wide scale. Unlike the existing work that mainly focuses on the statistical challenge, FedHybrid establishes a hierarchical hybrid training framework that effectively utilizes the fragmented and unbalanced data located on the participating devices on a wide scale. Specifically, FedHybrid consists of the following two core components, a global coordinator deployed on the central server and a local coordinator deployed on each participating device. The global coordinator organizes the participating devices into different groups through jointly considering the system heterogeneity and unbalanced training data in order to accelerate the overall training progress while guaranteeing the model performance. Within each group, a novel device-to-device (D2D) sequential training procedure is coordinated by the local coordinator to effectively utilize the fragmented and unbalanced training data in order to intelligently update the local models. At the same time, we provide the theoretical analysis of FedHybrid and conduct extensive experiments to evaluate its effectiveness. The results show that FedHybrid effectively improves model accuracy up to 27% and accelerates the whole training process by 20% on average.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Embedding of meshes on rotator graphs

Embedding of meshes on rotator graphs

引用

Midwest Symposium on Circuits and Systems (MWSCAS)

作者： S. Ponnuswamy V. Chaudhary Dept. of Electr. & Comput. Eng. Wayne State Univ. Detroit MI USA Parallel and Distributed Computing Laboratory Department of Electrical and Computer Engineering Wayne State University Detroit MI USA

A set of directed permutation graphs called rotator graphs were proposed as an alternative to the star and pancake graphs for multiprocessor interconnection networks. The rotator graphs have a smaller diameter than star and pancake graphs for the same number of nodes, while sharing the properties of star, pancake, and binary hypercubes like maximal fault tolerance, partitionability, etc. In this paper we develop a class of algorithms for recognizing undirected mesh structures in n-rotator graphs. The average dilation of the embeddings are very low as compared to the dilation of the embedding. These embeddings will be very useful for regular computations with bi-directional requirements, in addition to the irregular computations in n-rotator graphs. Most of the results presented here equally apply to another set of directed Cayley graphs, the cycle prefix digraphs.< >

关键词： Multiprocessor interconnection networks Hypercubes Bidirectional control Concurrent computing Fault tolerance Embedded computing Routing Application software Sorting Computer vision

来源：评论

学校读者我要写书评

暂无评论

A science gateway getting ready for serving the international molecular simulation community 2

A science gateway getting ready for serving the internationa...

引用

2012 EGI Community Forum / EMI 2nd Technical Conference, EGICF-EMITC 2012

作者： Gesing, Sandra Herres-Pawlis, Sonja Birkenheuer, Georg Brinkmann, André Grunzke, Richard Kacsuk, Peter Kohlbacher, Oliver Kozlovszky, Miklos Krüger, Jens Müller-Pfefferkorn, Ralph Schäfer, Patrick Steinke, Thomas Center for Bioinformatics Department of Computer Science University of Tübingen Sand 14 Tübingen72076 Germany Department of Chemistry Ludwig-Maximilians-University Munich Butenandtstr. 5-13 München81377 Germany Paderborn Center for Parallel Computing University of Paderborn Warburger Str. 100 Paderborn33089 Germany Johannes Gutenberg-University Mainz Mainz55099 Germany Center for Information Services and High Performance Computing Technische Universität Dresden Zellescher Weg 12-14 Germany Laboratory of Parallel and Distributed Systems MTA SZTAKI Kende Street 13-17 Budapest1111 Hungary Zuse Institute Berlin Takustraße 7 Berlin14195 Germany

The project MoSGrid (Molecular Simulation Grid) has been developing a web-based science gateway supporting the community with various services for quantum chemistry, molecular modeling, and docking. Users gain access to distributed computing infrastructures (DCIs) via intuitive user interfaces for sophisticated tools, specialized workflows, and distributed repositories. Currently, the MoSGrid community consists of about 120 users from a number of fields related to chemistry and bioinformatics located in Germany. However, the underlying security infrastructure is generally applicable and can be deployed in arbitrary projects. MoSGrid intends to address the international community by participating in the EU-projects SCI-BUS (Scientific gateway Based User Support) and ER-flow (Building an European Research Community through Interoperable Workflows and Data), and collaborating with the EU-project EDGI (European Desktop Grid Initiative). © Copyright owned by the author(s) under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike Licence.

关键词： Quantum chemistry

来源：评论

学校读者我要写书评

暂无评论

On implementation of a scalable wallet-size cluster computing system for multimedia applications

On implementation of a scalable wallet-size cluster computin...

引用

5th Pacific Rim Conference on Multimedia, PCM 2004

作者： Lee, Liang-Teh Li, Kuan-Ching Yang, Chao-Tung Tseng, Chia-Ying Liu, Kang-Yuan Hung, Chih-Hung Dept. of Computer Science and Engineering Tatung University Taipei104 Taiwan Parallel and Distributed Processing Center Dept. of Computer Science and Information Management Providence University Taichung433 Taiwan High Performance Computing Laboratory Dept. of Computer Science and Information Engineering Tunghai University Taichung407 Taiwan Networks and Multimedia Institute Institute for Information Industry Taipei106 Taiwan

ISBN: (纸本)9783540239857

In order to obtain a low-cost, low-power consumption and small sized digital home processing center, especially for multimedia applications, we propose a scalable wallet-size cluster computing system in this paper. Such system uses low-cost, low-power consumption and small sized microprocessors as computing nodes to form a cluster system. The main reason why we are interested to investigate and build such system is to increase the computing power and to accelerate the H.264 video processing. We analyzed the H.264 decoder scheme and distributed the most heavy-load computations in the decoder scheme to each computing node (or microprocessor) of the cluster system for the processing of signals, in parallel. To run multimedia applications in this cluster system, an efficient real-time video processing is implemented. © Springer-Verlag Berlin Heidelberg 2004.

关键词： computing power

来源：评论

学校读者我要写书评

暂无评论

DW-MLSR: Unsupervised Deformable Medical Image Registration Based on Dual-Window Attention and Multi-Latent Space

SSRN

引用

SSRN 2024年

作者： Huang, YuXuan Yin, MengXiao Zhong, Cheng Li, ZhiPan Yang, Feng School of Computer and Electronic Information Guangxi University Nanning530004 China Guangxi Key Laboratory of Multimedia Communications Network Technology Guangxi University Guangxi Nanning530004 China Key Laboratory of Parallel Distributed and Intelligent Computing of Guangxi Universities and Colleges Nanning530004 China

Thanks to the emergence of transformers and Vision Transformer (VIT), attention mechanisms have also been applied to medical image registration. However, the current attention mechanisms in medical image registration suffer from incomplete information interaction. Common unsupervised registration methods may be constrained by the imperfect medical image data with uncertainty and noise, leading to unstable registration results susceptible to *** address these evident drawbacks, this paper proposes a novel medical image registration method, namely, Registration Based on Dual-Window Attention and Multi-Latent Space(DW-MLSR). The Dual-Window attention mechanism introduced in this paper enables the model to comprehensively focus on crucial local regions in the image, thereby improving registration accuracy and robustness. Meanwhile, MLS, employed as a data augmentation method, enhances the model's generalization ability and robustness by learning latent image representations, providing richer image *** results demonstrate the superiority of the proposed medical image registration method over mainstream registration models. Comparative experiments indicate that our approach holds significant application potential in the field of medical image registration. © 2024, The Authors. All rights reserved.

关键词： Image enhancement

来源：评论

学校读者我要写书评

暂无评论

On the parallels between Paxos and Raft, and how to port optimizations (Extended Version)

arXiv

引用

arXiv 2019年

作者： Wang, Zhaoguo Zhao, Changgeng Mu, Shuai Chen, Haibo Li, Jinyang Shanghai Key Laboratory of Scalable Computing and Systems Shanghai Jiao Tong University Institute of Parallel and Distributed Systems Shanghai Jiao Tong University Department of Computer Science New York University Department of Computer Science Stony Brook University

In recent years, Raft has overtaken Paxos as the consensus algorithm of choice. [53] While many have pointed out similarities between the two protocols, no one has formally mapped out their relationships. In this paper, we show how Raft and Paxos are formally related despite their surface differences. Based on the formal mapping between the two protocols, we show how to automatically port a certain class of optimizations from Paxos to Raft with guaranteed correctness. As case studies, we port and evaluate two optimizations, Mencius and Paxos Quorum Lease to Raft. Copyright © 2019, The Authors. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

RoSAS: Deep Semi-supervised Anomaly Detection with Contamination-resilient Continuous Supervision

arXiv

引用

arXiv 2023年

作者： Xu, Hongzuo Wang, Yijie Pang, Guansong Jian, Songlei Liu, Ning Wang, Yongjun National Key Laboratory of Parallel and Distributed Computing China College of Computer National University of Defense Technology Hunan Changsha410073 China School of Computing and Information Systems Singapore Management University Singapore178902 Singapore College of Information and Communications National University of Defense Technology Hubei Wuhan430010 China

Semi-supervised anomaly detection methods leverage a few anomaly examples to yield drastically improved performance compared to unsupervised models. However, they still suffer from two limitations: 1) unlabeled anomalies (i.e., anomaly contamination) may mislead the learning process when all the unlabeled data are employed as inliers for model training;2) only discrete supervision information (such as binary or ordinal data labels) is exploited, which leads to suboptimal learning of anomaly scores that essentially take on a continuous distribution. Therefore, this paper proposes a novel semi-supervised anomaly detection method, which devises contamination-resilient continuous supervisory signals. Specifically, we propose a mass interpolation method to diffuse the abnormality of labeled anomalies, thereby creating new data samples labeled with continuous abnormal degrees. Meanwhile, the contaminated area can be covered by new data samples generated via combinations of data with correct labels. A feature learning-based objective is added to serve as an optimization constraint to regularize the network and further enhance the robustness w.r.t. anomaly contamination. Extensive experiments on 11 real-world datasets show that our approach significantly outperforms state-of-the-art competitors by 20%-30% in AUC-PR and obtains more robust and superior performance in settings with different anomaly contamination levels and varying numbers of labeled anomalies. The source code is available at https://***/xuhongzuo/rosas/. © 2023, CC BY.

关键词： Contamination

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：