检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Gao, Fei Hu, Ming Xie, Zhiyu Shi, Peichang Xie, Xiaofei Yi, Guodong Wang, Huaimin National Key Lab. of Parallel and Distributed Processing National University of Defense Technology Changsha China School of Computing and Information Systems Singapore Management University Singapore Xiangjiang Lab Changsha China

With advancements in AI infrastructure and Trusted Execution Environment (TEE) technology, Federated Learning as a Service (FLaaS) through Jointcloud computing (JCC) is promising to break through the resource constraints caused by heterogeneous edge devices in the traditional Federated Learning (FL) paradigm. Specifically, with the protection from TEE, data owners can achieve efficient model training with high-performance AI services in the cloud. By providing additional FL services, cloud service providers can achieve collab.rative learning among data owners. However, FLaaS still faces three challenges, i.e., i) low training performance caused by heterogeneous data among data owners, ii) high communication overhead among different clouds (i.e., data centers), and iii) lack of efficient resource scheduling strategies to balance training time and cost. To address these challenges, this paper presents a novel asynchronous FL approach named NebulaFL for collab.rative model training among multiple clouds. To address data heterogeneity issues, NebulaFL adopts a version control-based asynchronous FL training scheme in each data center to balance training time among data owners. To reduce communication overhead, NebulaFL adopts a decentralized model rotation mechanism to achieve effective knowledge sharing among data centers. To balance training time and cost, NebulaFL integrates a reward-guided strategy for data owners selection and resource scheduling. The experimental results demonstrate that, compared to the state-of-the-art FL methods, NebulaFL can achieve up to 5.71% accuracy improvement. In addition, NebulaFL can reduce up to 50% communication overhead and 61.94% costs under a target accuracy. Copyright © 2024, The Authors. All rights reserved.

关键词： Costs

来源：评论

学校读者我要写书评

暂无评论

Joint Power Splitting and Secure Beamforming Design in the Wireless-powered Untrusted Relay Networks

Joint Power Splitting and Secure Beamforming Design in the W...

引用

IEEE Global Communications Conference

作者： Mingxiong Zhao Suili Feng Xiangfeng Wang Meng Zhang Yuan Liu Hao Fu School of Electronic and Information Engineering South China University of Technology Shanghai Key Lab. of Trustworthy Computing Software Engineering Institute East China Normal Univ.

ISBN: (纸本)9781479959532

In this work, we maximize the secrecy rate of the wireless-powered untrusted amplify-and-forward relay networks by jointly designing power splitting (PS) ratio and relay beamforming with the proposed global optimal algorithm (GOA) and local optimal algorithm (LOA). To guarantee secure communication, the destination-based artificial noise is sent to degrade the reception of the untrusted relay, and it also becomes a new source of energy powering relay to forward the information with power splitting (PS) technique. Simulation result shows that LOA can achieve satisfactory secrecy rate performance compared with that of GOA, but with less computation time. It also manifests that both proposed algorithms outperform the benchmark method.

关键词： Relay networks power splitting TRIM47 gene Loa computational time Beamforming reception relay Privacy

来源：评论

学校读者我要写书评

暂无评论

Automatic adaptation of annotation standards: Chinese word segmentation and POS tagging: a case study 09

Automatic adaptation of annotation standards: Chinese word s...

引用

Proceedings of the Joint Conference of the 47th Annual Meeting of the ACL and the 4th International Joint Conference on Natural Language Processing of the AFNLP: Volume 1 - Volume 1

作者： Wenbin Jiang Liang Huang Qun Liu Key Lab. of Intelligent Information Processing Institute of Computing Technology Chinese Academy of Sciences Beijing China Google Research Charleston Rd. Mountain View CA

ISBN: (纸本)9781932432459

Manually annotated corpora are valuable but scarce resources, yet for many annotation tasks such as treebanking and sequence lab.ling there exist multiple corpora with different and incompatible annotation guidelines or standards. This seems to be a great waste of human efforts, and it would be nice to automatically adapt one annotation standard to another. We present a simple yet effective strategy that transfers knowledge from a differently annotated corpus to the corpus with desired annotation. We test the efficacy of this method in the context of Chinese word segmentation and part-of-speech tagging, where no segmentation and POS tagging standards are widely accepted due to the lack of morphology in Chinese. Experiments show that adaptation from the much larger People's Daily corpus to the smaller but more popular Penn Chinese Treebank results in significant improvements in both segmentation and tagging accuracies (with error reductions of 30.2% and 14%, respectively), which in turn helps improve Chinese parsing accuracy.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Efficiently Finding Top-K Items from Evolving Distributed Data Streams

Efficiently Finding Top-K Items from Evolving Distributed Da...

引用

International Conference on Semantics, Knowledge and Grid (SKG)

作者： Baoyuan Qi Gang Ma Zhongzhi Shi Wei Wang University of Chinese Academy of Sciences Beijing China Key Lab of Intelligent Information Processing Institute of Computing Technology Cas Beijing China Key Lab. of Intell. Inf. Process. Institute of Computing Technology Chinese Academy of Sciences Beijing CN Beijing Lexo Technologies Co. Ltd. Beijing China

ISBN: (纸本)9781479967162

The problem of efficiently finding top-k frequent items has attracted much attention in recent years. Storage constraints in the processing node and intrinsic evolving feature of the data streams are two main challenges. In this paper, we propose a method to tackle these two challenges based on space-saving and gossip-based algorithms respectively. Our method is implemented on SAMOA, a scalab.e advanced massive online analysis machine learning framework. The experimental results show its effectiveness and scalab.lity.

关键词： Peer-to-peer computing Data mining Distributed databases Protocols Machine learning algorithms Monitoring Radiation detectors

来源：评论

学校读者我要写书评

暂无评论

Accelerating deep unsupervised domain adaptation with transfer channel pruning

arXiv

引用

arXiv 2019年

作者： Yu, Chaohui Wang, Jindong Chen, Yiqiang Wu, Zijing Beijing Key Lab. of Mobile Computing and Pervasive Device Inst. of Computing Technology CAS Beijing China University of Chinese Academy of Sciences Beijing China Columbia University New YorkNY10024 United States

Deep unsupervised domain adaptation (UDA) has recently received increasing attention from researchers. However, existing methods are computationally intensive due to the computation cost of Convolutional Neural Networks (CNN) adopted by most work. To date, there is no effective network compression method for accelerating these models. In this paper, we propose a unified Transfer Channel Pruning (TCP) approach for accelerating UDA models. TCP is capable of compressing the deep UDA model by pruning less important channels while simultaneously learning transferable features by reducing the crossdomain distribution divergence. Therefore, it reduces the impact of negative transfer and maintains competitive performance on the target task. To the best of our knowledge, TCP is the first approach that aims at accelerating deep UDA models. TCP is validated on two benchmark datasets Office-31 and ImageCLEFDA with two common backbone networks VGG16 and ResNet50. Experimental results demonstrate that TCP achieves comparable or better classification accuracy than other comparison methods while significantly reducing the computational cost. To be more specific, in VGG16, we get even higher accuracy after pruning 26% floating point operations (FLOPs);in ResNet50, we also get higher accuracy on half of the tasks after pruning 12% FLOPs. We hope that TCP will open a new door for future research on accelerating transfer learning models. Copyright © 2019, The Authors. All rights reserved.

关键词： Digital arithmetic

来源：评论

学校读者我要写书评

暂无评论

Ranking Sentences in Scientific Literatures

Ranking Sentences in Scientific Literatures

引用

International Conference on Semantics, Knowledge and Grid (SKG)

作者： Jiao Tian Mengyun Cao Jin Liu Xiaoping Sun Zhuge Hai University of Chinese Academy of Sciences Beijing China Knowledge Grid Group CAS Key Lab of Intelligent Information Processing Institute of Computing Technology Chinese Academy of Sciences Beijing China Wuhan University CAS Key Lab. of Intell. Inf. Process. Institute of Computing Technology Chinese Academy of Sciences Beijing CN

Sentence ranking is one of the most important research issues in text analysis. It can be used in extracted textual summarization and information retrieval. Graph-based methods are a common way of ranking and extracting sentences. In graph based methods, sentences are nodes of graph and edges are built based on the sentence similarities or on sentence co-occurrence relationship. PageRank style algorithms can be applied to get sentence ranks. In this paper, we focus on how to rank sentences in a single scientific paper. A scientific literature has more structural information than general texts and this structural information has not been fully explored yet in graph based ranking models. We investigated several different methods that used structural information such as paragraph and section information to construct a heterogeneous graph for ranking sentences. We conducted experiments on these methods to compare the results on sentence ranking. It shows that structural information can help identify more representative sentences.

关键词： Semantics Algorithm design and analysis Web pages Data mining Iterative methods Current measurement

来源：评论

学校读者我要写书评

暂无评论

Relation-Guided Representation Learning

arXiv

引用

arXiv 2020年

作者： Kang, Zhao Lu, Xiao Liang, Jian Bai, Kun Xu, Zenglin School of Computer Science and Engineering University of Electronic Science and Technology of China Sichuan China Trusted Cloud Computing and Big Data Key Laboratory of Sichuan Province Cloud and Smart Industries Group Tencent Beijing China School of Computer Science and Technology Harbin Institute of Technology Shenzhen China Center for Artificial Intelligence Peng Cheng Lab Shenzhen China

Deep auto-encoders (DAEs) have achieved great success in learning data representations via the powerful representability of neural networks. But most DAEs only focus on the most dominant structures which are able to reconstruct the data from a latent space and neglect rich latent structural information. In this work, we propose a new representation learning method that explicitly models and leverages sample relations, which in turn is used as supervision to guide the representation learning. Different from previous work, our framework well preserves the relations between samples. Since the prediction of pairwise relations themselves is a fundamental problem, our model adaptively learns them from data. This provides much flexibility to encode real data manifold. The important role of relation and representation learning is evaluated on the clustering task. Extensive experiments on benchmark data sets demonstrate the superiority of our approach. By seeking to embed samples into subspace, we further show that our method can address the large-scale and out-of-sample problem. Our source code is publicly availab.e at: https://***/nbShawnLu/RGRL. Copyright © 2020, The Authors. All rights reserved.

关键词： Clustering algorithms

来源：评论

学校读者我要写书评

暂无评论

Grrln: Gated Recurrent Residual Learning Networks for Code Clone Detection

SSRN

引用

SSRN 2022年

作者： Zhang, Xiangping Xiao, Qiaoxiang Hu, Haize Liu, Jianxun Hunan Provincial Key Lab. for Services Computing and Novel Software Technology Hunan University of Science and Technology Hunan Xiangtan411100 China School of Computer Science and Technology Hunan University of Science and Technology Hunan Xiangtan411100 China

Code clone detection is a critical problem in software development and maintenance domains. It aims to identify functionally identical or similar code fragments within an application. Existing works formulate the code clone detection task as a binary classification problem which is determine the code pairs are clone or not with a pre-defined threshold. In reality, there are various types of code clone subject to the degree of how a pair of code fragments are similar to each other. To investigate the effect of different code clone detection manner on the clone detection result, we propose Gated Recurrent Residual Learning Networks (GRRLN), a novel neural network model for code clone detection. To train GRRLN, we first represent each code fragment as a statement-level tree sequence derived from the whole AST. Then, a gated recurrent neural network with residual connections is adopted to fully extract semantics of all individual statement trees together with their dependency relationships across the input statement sequence. Finally, the output representations of code fragments by GRRLN are used for similarity calculation and clone detection. We evaluate GRRLN using two real-world datasets for code clone detection and clone type classification. Experiments show that GRRLN achieves generally better performance than state-of-the-art methods and meanwhile needs significant less time and memory costs. © 2022, The Authors. All rights reserved.

关键词： Recurrent neural networks

来源：评论

学校读者我要写书评

暂无评论

G-MAP: General Memory-Augmented Pre-trained Language Model for Domain Tasks

arXiv

引用

arXiv 2022年

作者： Wan, Zhongwei Yin, Yichun Zhang, Wei Shi, Jiaxin Shang, Lifeng Chen, Guangyong Jiang, Xin Liu, Qun Guangdong Provincial Key Laboratory of Computer Vision and Virtual Reality Technology Shenzhen Institute of Advanced Technology Chinese Academy of Science China University of Chinese Academy of Sciences China Huawei Noah's Ark Lab Hong Kong Huawei Cloud Computing Zhejiang Lab China

Recently, domain-specific PLMs have been proposed to boost the task performance of specific domains (e.g., biomedical and computer science) by continuing to pre-train general PLMs with domain-specific corpora. However, this Domain-Adaptive Pre-Training (DAPT;Gururangan et al. (2020)) tends to forget the previous general knowledge acquired by general PLMs, which leads to a catastrophic forgetting phenomenon and sub-optimal performance. To alleviate this problem, we propose a new framework of General Memory- Augmented Pre-trained Language Model (GMAP), which augments the domain-specific PLM by a memory representation built from the frozen general PLM without losing any general knowledge. Specifically, we propose a new memory-augmented layer, and based on it, different augmented strategies are explored to build the memory representation and then adaptively fuse it into the domain-specific PLM. We demonstrate the effectiveness of G-MAP on various domains (biomedical and computer science publications, news, and reviews) and different kinds (text classification, QA, NER) of tasks, and the extensive results show that the proposed G-MAP1 can achieve SOTA results on all tasks. © 2022, CC BY.

关键词： Computational linguistics

来源：评论

学校读者我要写书评

暂无评论

Exploring the Benefits of Introducing Network Coding into Named Data Networking

Exploring the Benefits of Introducing Network Coding into Na...

引用

IEEE International Conference on Computational Science and Engineering, CSE

作者： Kai Lei Tao Chen Chen Peng Zhi Tan Shenzhen Key Lab for Cloud Computing Technology & Applications (SPCCTA) School of Electronics and Computer Engineering (SECE) Peking University Shenzhen P.R. China GuangDong Poya Information & Technology Co. LTD Guangzhou P.R. China

ISBN: (纸本)9781479979820

In recent years, the focus to optimize network transmission efficiency has evolved to adopt methods that let those intermediate data transferring nodes get involved with routing, forwarding and caching. In other words, the new network architecture designs become in favor of hop-to-hop model, instead of traditional TCP-like end-to-end model. Named data networking is a promising future internet data oriented architecture which uses names instead of addresses and exchanges or forwards interest/data pair packets at each node along the path to route data for delivery. And meanwhile Network coding (NC) is a content oriented and effective method to reduce redundancy, increase network throughput and improve robustness. Nonetheless, due to NDN's current preliminary research, less research has combined these two technologies together. This paper presents some new thoughts to study on the benefits brought by integrating network coding to NDN, which can effectively improve network utilization, strengthen caching privacy, and also promote development of the NDN architecture itself.

关键词： Network coding Security Peer-to-peer computing Communication networks Encoding Bandwidth Internet

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：