检索结果-内蒙古大学图书馆

IEEE International Conference on Multimedia and Expo (ICME)

作者： He Zhu Yang Chen Guyue Hu Shan Yu Brainnetome Center National Laboratory of Pattern Recognition (NLPR) Institute of Automation Chinese Academy of Sciences(CASIA) School of Future Technology University of Chinese Academy of Sciences(UCAS) School of Computer Science and Engineering Nanyang Technological University School of Artificial Intelligence University of Chinese Academy of Sciences(UCAS)

Recent representation learning approaches mainly fall into two paradigms: contrastive learning (CL) and masked image modeling (MIM). Combining these two methods may boost the performance, but its learning process still heavily depends on the random masking strategy. We conjecture that the random masking may hinder learning the comprehensive relationship between concept and visual patches. To overcome these limitations, we propose an information-density masking (IDM) strategy for general visual transformers. Specifically, the IDM mask out the visual patches according to their activation values of attention maps. To obtain the attention maps before the reconstruction, a self-supervised training framework CAMAE is further proposed. In addition, in order to reduce the redundancy among different attention maps, we introduce a pattern-learning balance (PLB) sampling to adaptively adjust the learning progress in different attention spaces. Extensive experiments indicate that our method efficiently retains more comprehensive visual characteristics and achieves state-of-the-art performance.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Dynamic Identity-Based Cryptography using RSA

Dynamic Identity-Based Cryptography using RSA

引用

International Conference on Computing and Networking technology (ICCNT)

作者： Gopala Krishnan M.P. Karnik S. Hemelatha Simranjeet Nanda P Allirani Amit Kansal School of Mangement-PG Management JAIN (Deemed to be University) Bangalore Karnataka India Computer Engineering Vishwakarma Institute of Technology Pune India Department of Computer Science Engineering Karpagam Academy of Higher Education Coimbatore India Centre of Research Impact and Outcome Chitkara University Rajpura Punjab India Department of Information Technology Prince Shri Venkateshwara Padmavathy Engineering College Chennai India Quantum University Research Center Quantum University

ISBN: (数字)9798350370249

ISBN: (纸本)9798350370270

Dynamic identification-based cryptography is a rising shape of cryptographic era that makes use of best an identity in place of a pair of public and personal keys to encrypt facts. This technology is based totally at the RSA cryptosystem and makes use of the identity of a person to generate an ephemeral public key for each verbal exchange session. By using utilizing the identity of the person, dynamic identity-based totally cryptography can decrease the quantity of computational resources had to authenticate and encrypt messages for the duration of communiqué. Furthermore, this technology lets in for more potent encryption on the grounds that a person’s identification changes over time so that the public key ought to be re-generated for every communiqué consultation. This method of encryption has the ability for use in more than one program, which includes mystery communication between government agencies and people, monetary applications and e-trade transactions.

关键词： Identity-based encryption Protocols Shape Government Buildings Public key Dynamic scheduling Encryption

来源：评论

学校读者我要写书评

暂无评论

Learning from models beyond fine-tuning

arXiv

引用

arXiv 2023年

作者： Zheng, Hongling Shen, Li Tang, Anke Luo, Yong Hu, Han Du, Bo Wen, Yonggang Tao, Dacheng School of Computer Science School of Artificial Intelligence National Engineering Research Center for Multimedia Software Hubei Key Laboratory of Multimedia and Network Communication Engineering Wuhan University China School of Cyber Science and Technology Shenzhen Campus of Sun Yat-sen University JD Explore Academy China Beijing Institute of Technology China Nanyang Technological University Singapore

Foundation models (FM) have demonstrated remarkable performance across a wide range of tasks (especially in the fields of natural language processing and computer vision), primarily attributed to their ability to comprehend instructions and access extensive, high-quality data. This not only showcases their current effectiveness but also sets a promising trajectory towards the development of artificial general intelligence. Unfortunately, due to multiple constraints, the raw data of the model used for large model training are often inaccessible, so the use of end-to-end models for downstream tasks has become a new research trend, which we call Learn From Model (LFM) in this article. LFM focuses on the research, modification, and design of FM based on the model interface, so as to better understand the model structure and weights (in a black box environment), and to generalize the model to downstream tasks. The study of LFM techniques can be broadly categorized into five major areas: model tuning, model distillation, model reuse, meta learning and model editing. Each category encompasses a repertoire of methods and strategies that aim to enhance the capabilities and performance of FM. This paper gives a comprehensive review of the current methods based on FM from the perspective of LFM, in order to help readers better understand the current research status and ideas. To conclude, we summarize the survey by highlighting several critical areas for future exploration and addressing open issues that require further attention from the research community. The relevant papers we investigated in this article can be accessed at https://***/ruthless-man/Awesome-Learn-from-Model. © 2023, CC BY.

关键词： Frequency modulation

来源：评论

学校读者我要写书评

暂无评论

Document-level DDI relation extraction with document-entity embedding

Document-level DDI relation extraction with document-entity ...

引用

2021 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2021

作者： Mingliang, Dou Jijun, Tang Fei, Guo Tianjin University College of Intelligence and Computing Tianjin China Chinese Academy of Sciences Shenzhen Institute of Advanced Technology Shenzhen China Central South University School of Computer Science and Engineering Changsha China

ISBN: (纸本)9781665401265

DDI is an important part of drug-related research and pharmacovigilance. Extracting DDI information from scientific literature has become a low-cost and highly reliable way. Currently, existing works are all sentence-level DDI relation extraction. In fact, the entity relationship is often expressed by multiple sentences. Moreover, the sentence-level DDI relation extraction also causes a large amount of redundancy in the whole dataset with increasing in negative instance data. In this study, we propose a document-level DDI relation extraction method based on document-entity embedding. Our method performs special processing on the DDI Extraction 2013 for the first time, in order to calculate document-level relation extraction. For obtaining document-level entity information, we propose a document-entity embedding method to integrate the information of all same drugs in the same article. The experimental results show that the processing of DDI Extraction 2013 dataset is reasonable. In addition, the proposed method has achieved good performance on document-level DDI dataset, and the best F1 score is 62.51%. This is the first time that DDI Extraction 2013 has been processed into a document-level dataset, and document-level DDI relation extraction has been realized. © 2021 IEEE.

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

Automatic Topic Classification And Document Clustering Using Lda Based Machine Learning Techniques

Automatic Topic Classification And Document Clustering Using...

引用

Communications and Information Technologies (GCCIT), Global Conference on

作者： K. Yuvaraj Nipun Setia Sangeetha S Prakhar Goyal Thomas P K Vishal A. Meshram Department of Computer Science Karpagam Academy of Higher Education Coimbatore Centre of Research Impact and Outcome Chitkara University Rajpura Punjab India Department of Science and Humanities Prince Shri Venkateshwara Padmavathy Engineering College Chennai Quantum University Research Center Quantum University Roorkee Uttarakhand India Adjunct Faculty School of Management - PG Management JAIN (Deemed to be University) Bangalore Karnataka Department of Computer Engineering Vishwakarma Institute of Technology Pune India

ISBN: (数字)9798350388916

ISBN: (纸本)9798350388923

Since more and more information is being stored in digital formats (such as databases, journals, libraries, and papers), it is crucial that this information be put to good use. However, it might be challenging to extract meaningful insights from the data because of its poor organisation, ambiguity, and lack of precision. Data may be correctly categorised to allow for separate research on different types of data, which will help with this problem. Because of this, researchers in disciplines like mathematics and computer science have been trying to figure out how to fix the issue. Material mining is a branch of machine learning that helps cope with unstructured and disorganised data by sifting through it for useful insights. Text mining is a subfield of data mining that focuses on discovering patterns and extracting structured attributes from unstructured text. To successfully analyse such data, supervised learning is combined with unsupervised learning. It has become common practise to employ Latent Dirichlet Allocation (LDA), a probabilistic approach to topic modelling, when classifying documents. As an unsupervised method, however, traditional LDA's findings may shift depending on which articles come first in the analysis. In this study, we provide a comprehensive strategy for avoiding the order effect as well as the unsupervised probabilistic character of the problem. The proposed framework achieved a success rate of 95.56% & 95.23% when evaluated on two separate datasets, demonstrating its better classification capabilities. Furthermore, the subjects that were extracted from the classified clusters are highly relevant to the domain classifications.

关键词： Text mining Machine learning algorithms Databases Machine learning Feature extraction Probabilistic logic Mathematics Libraries Data mining Unsupervised learning

来源：评论

学校读者我要写书评

暂无评论

Data-consistent Unsupervised Diffusion Model for Metal Artifact Reduction

Data-consistent Unsupervised Diffusion Model for Metal Artif...

引用

IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

作者： Zhan Tong Zhan Wu Yang Yang Weilong Mao Shijie Wang Yinsheng Li Yang Chen Laboratory of Image Science and Technology Southeast University Nanjing China Key Laboratory of Computer Network and Information Integration Southeast University Ministry of Education Nanjing China Research Center for Medical Artificial Intelligence Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences Shenzhen China Jiangsu Provincial Joint International Research Laboratory of Medical Information Processing the Laboratory of Image Science and Technology School of Computer Science and Engineering and the Key Laboratory of New Generation Artificial Intelligence Technology and Its Interdisciplinary Applications Nanjing China

Computed Tomography (CT) is an imaging technique widely used in clinical diagnosis. However, high-attenuation metallic implants result in the obstruction of low-energy Xrays and further lead to metal artifacts in the reconstructed CT images. Deep supervised model-based metal artifact reduction(MAR) approaches are limited in clinical applications due to the difficulty in obtaining paired artifact-affected and artifactfree data. Furthermore, these model-based methods lack the consideration of data consistency in the sinogram-domain to perform exact metal trace inpainting. To address these challenges, we propose a Data-consistent unsupErVised diffusiOn model for meTal artifact rEDuction, called DEVOTED-Net. First, DEVOTED-Net leverages prior knowledge to guide the conditional diffusion model for fine-grained metal trace inpainting. Second, an unsupervised MAR framework is designed in the reverse process for the unknown metal traces restoration in the sinogram domain. Third, to further enhance the sinogram-domain data consistency, physics-based consistency constraint loss including conjugateray consistency loss and accumulation-ray consistency loss is designed. Extensive experiments are carried out to verify the performance of our algorithm on the publicly available dataset and clinical experimental dataset. This efficient, accurate, and reliable MAR approach holds great potential in clinics.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Pyscribe - Learning to Describe Python Code

SSRN

引用

SSRN 2022年

作者： Guo, Juncai Liu, Jin Wan, Yao Zhao, Yanjie Li, Li Liu, Kui Klein, Jacques Bissyandé, Tegawendé F. School of Computer Science Wuhan University Wuhan China Key Laboratory of Network Assessment Technology Institute of Information Engineering Chinese Academy of Sciences Beijing China School of Computer Science and Technology Huazhong University of Science and Technology Wuhan China Faculty of Information Technology Monash University VIC Australia Huawei Software Engineering Application Technology Lab Hangzhou China SnT Centre University of Luxembourg 1855 Luxembourg

Code comment generation, which aims to describe the functionality of source code in natural language, plays an important role in automatic software development. Currently, several structural neural networks have been exploited to preserve the syntax structure of source code based on Abstract Syntax Trees (ASTs). However, they can not well capture both the long-distance and local relations between nodes while retaining the overall structural information of AST. To mitigate this problem, we present to the community a prototype tool called PYSCRIBE, which expands the Transformer model to a new encoder-decoder architecture to incorporate triplet positions into the node-level and edge-level features of AST towards automatically generating comments for Python code. To the best of our knowledge, this paper makes the first attempt to consider the edges of AST as an explicit component for better code representation, as to generate higher qualified comments. By specifying triplet positions for each node and edge, the overall structural information can be well preserved in the learning process. In the decoder, we perform a two-stage decoding process over the extracted node and edge features to generate comments. To evaluate the effectiveness of PYSCRIBE, we resort to a large dataset of code-comment pairs by mining Jupyter notebooks from GitHub, for which we have made it publicly available to support further studies. The evaluation results show that PYSCRIBE is indeed effective, outperforming the state-of-the-art by achieving an average BLEU score (i.e., av-BLEU) of ≈0.28. © 2022, The Authors. All rights reserved.

关键词： Large dataset

来源：评论

学校读者我要写书评

暂无评论

Transformation from 5G N/W to 6G N/W Through technology Inclusion

Transformation from 5G N/W to 6G N/W Through Technology Incl...

引用

Advance Computing and Innovative Technologies in engineering (ICACITE), International Conference on

作者： Rohit Vaish V. Diana Earshia M Jeba Paulin Hassan M. Al-Jawahry B Maheshwari T. Mohanraj Department of Electronics & Communication Engineering IES College Of Technology Bhopal M. P India Vel Tech Rangarajan Dr Sagunthala R&D Institute of Science and Technology (Deemed To Be University) Avadi Chennai TamilNadu India Department of Electronics and Communication Engineering Nehru Institute of Technology T.M.Palayam Tamil Nadu India Department of computers Techniques engineering College of technical engineering The Islamic University Najaf Iraq Department of Electronics and Communication Engineering Prince Shri Venkateshwara Padmavathy Engineering College TN Department of Computer Science Engineering Karpagam Academy of Higher Education Coimbatore TN

ISBN: (数字)9798350360165

ISBN: (纸本)9798350360172

For us, figuring out where we are or where our resources are has always been important. While technology has greatly benefited this quest, real-time interior locating remains a daunting problem, from conventional instruments like sextants to contemporary outside GPS devices. With the introduction of 5G mobile networks, network-based location became a practical option. With the help of sophisticated radio technologies, decreased end-to-end latency, specialist control protocols, and increased processing power at the network edge, 5G’s full promise for indoor and outdoor placement could now be realized. This study provides a thorough introduction to network-based positioning, examining advanced machine-learning-supported solutions as well as underlying ideas. A detailed comparison of learning methods used in network-based positioning is a significant addition. The article provides a forward-looking strategy to positioning using these advanced networks, as we expect the progression to 6G networks. Various application examples from different sectors are offered to highlight the topic’s practical ramifications, with a focus on industrial industry vehicular applications. The goal of the material is to be both insightful and educational, serving as a useful tool for comprehending and managing the rapidly changing field of location-based technology.

关键词： 6G mobile communication Wireless communication Protocols 5G mobile communication Reviews Process control Standardization

来源：评论

学校读者我要写书评

暂无评论

Streaming Algorithms for Graph k-Matching with Optimal or Near-Optimal Update Time 32

Streaming Algorithms for Graph k-Matching with Optimal or Ne...

引用

32nd International Symposium on Algorithms and Computation, ISAAC 2021

作者： Chen, Jianer Huang, Qin Kanj, Iyad Li, Qian Xia, Ge Department of Computer Science and Engineering Texas A&M University College StationTX United States School of Computing DePaul University ChicagoIL United States Institute of Computing Technology Chinese Academy of Sciences Beijing China Department of Computer Science Lafayette College EastonPA United States

ISBN: (纸本)9783959772143

We present streaming algorithms for the graph k-matching problem in both the insert-only and dynamic models. Our algorithms, while keeping the space complexity matching the best known upper bound, have optimal or near-optimal update time, significantly improving on previous results. More specifically, for the insert-only streaming model, we present a one-pass randomized algorithm that runs in optimal O(k2) space and has optimal O(1) update time, and that, w.h.p. (with high probability), computes a maximum weighted k-matching of a weighted graph. Previously, the best upper bound on the update time was O(log k), which was achieved by a deterministic streaming algorithm that however only works for unweighted graphs [16]. For the dynamic streaming model, we present a one-pass randomized algorithm that, w.h.p., computes a maximum weighted k-matching of a weighted graph in Õ(Wk2) space1 and with Õ(1) update time, where W is the number of distinct edge weights. Again the update time of our algorithm improves the previous best upper bound Õ(k2) [7]. Moreover, we prove that in the dynamic streaming model, any randomized streaming algorithm for the problem requires k2 · Ω(W(log W + 1)) bits of space. Hence, both the space and update-time complexities achieved by our algorithm in the dynamic model are near-optimal. A streaming approximation algorithm for k-matching is also presented, whose space complexity matches the best known upper bound with a significantly improved update time. © Jianer Chen, Qin Huang, Iyad Kanj, Qian Li, and Ge Xia.

关键词： Dynamic models

来源：评论

学校读者我要写书评

暂无评论

Generation of Dynamic Authorization Using Cryptography For Private-Keyless Data Sharing

SSRN

引用

SSRN 2023年

作者： Lu, Hai Zhu, Yan Yau, Stephen S. Zhu, Guizhen Lv, Kewei School of Computer and Communication Engineering University of Science and Technology Beijing Beijing100083 China School of Computing Informatics and Decision Systems Engineering Arizona State University TempeAZ85281 United States Data and Communication Science Technology Research Institute Beijing100191 China School of Cyber Security University of Chinese Academy of Sciences Beijing100093 China

Private-keyless access is emerging as an inevitable trend in the evolution of future information systems. It helps simplify key management, enhance security, and increase flexibility of access authorization. Hereby, we introduce a novel security model, Cryptographic Resource Enabled Infrastructure (CREI), to support private-keyless access and policy-based data custody. Firstly, in the data custody phase, data resource and its session key are uploaded in ciphertext form to CREI, and the access authorization complies with multiple policies designated by the provider. Secondly, in the dynamic authorization phase, traditional keys or keycards are replaced with real-time attribute credentials issued by multiple authorities. CREI makes dynamically cryptographic policy decision according to the credentials and the cryptographic policy (cryptopolicy). Technically, we design the Credentials-Driven Cryptosystem (CDC) over ideal lattice to build anti-quantum confidence for CREI. In CDC, preimage Gaussian sampler is adopted to sample short vector as real-time attribute credential instead of user's private key. Small Policy Matrix (SPM) is implemented to convert monotone policies into optimized cryptopolicies in order to reduce accumulated error. The security analysis confirms that the credentials, cryptopolicy and its decision result are existentially unforgeable, ensuring the semantic security of the whole CREI. Experimental results indicate CREI has lower storage and computational costs than existing schemes. Meanwhile, despite continuous improvements to the access policies as system evolves, CREI can simply select an appropriate policy among multiple candidates to make authorization decisions without having to re-encrypt data. © 2023, The Authors. All rights reserved.

关键词： Authorization

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：