检索结果-内蒙古大学图书馆

Proceedings of the 60th Annual ACM/IEEE Design Automation Conference

作者： Hao Qi Yu Zhang Ligang He Kang Luo Jun Huang Haoyu Lu Jin Zhao Hai Jin National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Cluster and Grid Computing Lab School of Computer Science and Technology Huazhong University of Science and Technology China National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Cluster and Grid Computing Lab School of Computer Science and Technology Huazhong University of Science and Technology China and Zhejiang-HUST Joint Research Center for Graph Processing Zhejiang Lab China Department of Computer Science University of Warwick United Kingdom

ISBN: (纸本)9798350323481

Streaming Graph Pattern Mining (GPM) has been widely used in many application fields. However, the existing streaming GPM solution suffers from many unnecessary explorations and isomorphism tests, while the existing static GPM ones require many repetitive operations to compute the full graph. In this paper, we propose a pattern-aware incremental execution approach and design the first streaming GPM accelerator called PSMiner, which integrates multiple optimizations to reduce redundant computation and improve computing efficiency. We have conducted extensive experiments. The results show that compared with the state-of-the-art software and hardware solutions, PSMiner achieves the average speedups of 770.9× and 60.4×, respectively.

关键词：

来源：评论

学校读者我要写书评

暂无评论

On statistical rates and provably efficient criteria of latent diffusion transformers (DiTs) 24

On statistical rates and provably efficient criteria of late...

引用

Proceedings of the 38th International Conference on Neural Information Processing Systems

作者： Jerry Yao-Chieh Hu Weimin Wu Zhuoru Li Sophia Pi Zhao Song Han Liu Center for Foundation Models and Generative AI and Department of Computer Science Northwestern University Evanston IL Department of Statistics and Data Science Northwestern University Evanston IL Department of Computer Science Northwestern University Evanston IL Simons Institute for the Theory of Computing UC Berkeley Berkeley CA Center for Foundation Models and Generative AI and Department of Computer Science and Department of Statistics and Data Science Northwestern University Evanston IL

ISBN: (纸本)9798331314385

We investigate the statistical and computational limits of latent Diffusion Transformers (DiTs) under the low-dimensional linear latent space assumption. Statistically, we study the universal approximation and sample complexity of the DiTs score function, as well as the distribution recovery property of the initial data. Specifically, under mild data assumptions, we derive an approximation error bound for the score network of latent DiTs, which is sub-linear in the latent space dimension. Additionally, we derive the corresponding sample complexity bound and show that the data distribution generated from the estimated score function converges toward a proximate area of the original one. Computationally, we characterize the hardness of both forward inference and backward computation of latent DiTs, assuming the Strong Exponential Time Hypothesis (SETH). For forward inference, we identify efficient criteria for all possible latent DiTs inference algorithms and showcase our theory by pushing the efficiency toward almost-linear time inference. For backward computation, we leverage the low-rank structure within the gradient computation of DiTs training for possible algorithmic speedup. Specifically, we show that such speedup achieves almost-linear time latent DiTs training by casting the DiTs gradient as a series of chained low-rank approximations with bounded error. Under the low-dimensional assumption, we show that the statistical rates and the computational efficiency are all dominated by the dimension of the subspace, suggesting that latent DiTs have the potential to bypass the challenges associated with the high dimensionality of initial data.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Exploring transfer learning for Deep NLP systems on rarely annotated languages

arXiv

引用

arXiv 2024年

作者： Yadav, Dipendra Strauß, Tobias Yordanova, Kristina University of Rostock Department of Computer Science and Electrical Engineering Institute of Visual & Analytic Computing Mobile Multimedia Information Systems Albert-Einstein-Str. 22 Rostock18059 Germany

Natural language processing has seen exponential growth since deep learning has been used for various tasks. Deep learning has replaced the old rule-based methods by providing better performance on almost all the tasks in NLP, as deep learning basically looks into the hidden patterns and structures in the data which were sometimes not detected and at other-times overlooked by the rule-based systems. The research and development in NLP have been however focused on some handful languages of the world given their financial importance and a huge number of speakers. This has left out most of the other languages unexplored. The other reason for the lack of research in these languages is the unavailability of properly annotated and the required amount of datasets which is very important for training deep learning neural networks. The situation looks grim. When we look at the world language tree, thankfully we see that most of the unexplored languages have other well-explored languages nearby which basically means that both the languages are kind of similar. So if we could somehow adapt research and development from one language to another, it could be really helpful for both the languages. The goal of this thesis work is to explore transfer learning for the task of Part-of-speech(POS) tagging between Hindi and Nepali languages which are part of the Indo-Aryan language family with very high similarity. We have tried to explore the possibility of jointly training a model for the task of POS tagging in both Hindi and Nepali language and see if it helps in improving the performance of the model. We also try to explore if multitask learning in the Hindi language can be helpful for the task of POS tagging with auxiliary tasks of the gender tagging and singularity/plurality tagging. The deep learning architecture used for this work is BLSTM-CNN-CRF. The model is trained with monolingual word embeddings, vector mapped embeddings, and also with jointly trained Hindi-Nepali word emb

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

Digital Twin, Metaverse and Smart Cities in a Race to the Future

Digital Twin, Metaverse and Smart Cities in a Race to the Fu...

引用

Information Technology (ACIT)

作者： Hijab e Zainab Narmeen Zakaria Bawanay Department of Computer Science and Software Engineering Center for Computing Research Jinnah University for Women Karachi Pakistan

In the dynamic landscape of urban development, the synergy of the Metaverse, Digital Twins, and Smart Cities emerges as a potent catalyst for transformation. The Metaverse, a vibrant virtual space, collaborates with cutting-edge technologies, including Virtual Reality (VR), Augmented Reality (AR), the Internet of Things (IoT), and Artificial Intelligence (AI), crafting immersive experiences. Simultaneously, Digital Twins, serving as virtual replicas of physical entities, instigating an industrial revolution, while Smart Cities harness advanced technologies for sustainable urban living. This paper explores the applications of converging these technologies across diverse industries, showcasing how Digital Twins reshape conventional practices and amplify the concept of smart cities through the immersive realm of the Metaverse. However, the exploration unveils challenges, particularly in data privacy, interoperability, and ethical considerations. The study advocates for a balanced and informed approach, emphasizing that the symbiotic relationship between Digital Twins and the Metaverse should steer the evolution of Smart Cities toward security, innovation, and interconnected urban spaces.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Fault Tolerance and Cost Effectiveness Analysis of Low-Powered 3D-TESH Network 4

Fault Tolerance and Cost Effectiveness Analysis of Low-Power...

引用

4th International Conference on Sustainable Technologies for Industry 4.0, STI 2022

作者： Faisal, Faiz Al Hafizur Rahman, M.M. Inoguchi, Yasushi Green University of Bangladesh Department of Computer Science and Engineering Dhaka Bangladesh Ccsit King Faisal University Department of Computer Networks & Communications Saudi Arabia Jaist Research Center for Advanced Computing Infrastructure Nomi Ishikawa Japan

ISBN: (纸本)9781665490450

Exa-scale supercomputing is the prime target for the supercomputers. To reach at that point, supercomputers require redesigning the interconnection networks to support high network scalability, suitable network performance, low cost-performance ratio, low power usage, high fault-tolerance, low network congestion and high throughput with low latency. Among the thousand of interconnection networks, one of the most capable one is 3D-TESH, is a hierarchical interconnection network (HIN) interconnected with multiple basic modules (basic modules are basically treated as single on-chip network), where each on-chip network is similar to a 3Dmesh network and off-chip network are built using the 2Dtorus connections. Some static network parameters as well as dynamic communication performance (on-chip level) had been studied earlier for 3DTESH, which ensured the high significance of this network in comparing with others. Moreover, in case of power consumption, it requires about 41% less router power than 5Dtorus at the on-chip level. In this paper, we like to evaluate some other important static parameters like- packing density, fault tolerance, wiring complexity, cost effectiveness factor and etc. Here, we found that 3D-TESH yields much better results than network like - 3Dmesh, 2Dmesh, 2Dtorus, TESH and even the 3Dtorus. © 2022 IEEE.

关键词： Cost effectiveness

来源：评论

学校读者我要写书评

暂无评论

A Blockchain based Framework for Verifying Drug Authenticity

A Blockchain based Framework for Verifying Drug Authenticity

引用

Information Technology (ACIT)

作者： Soomaiya Hamid Narmeen Zakaria Bawany Kiran Fatima Department of Computer Science and Software Engineering Center for Computing Research Jinnah University for Women Karachi Pakistan

Counterfeit drugs are fake medicines that are potentially harmful for health. Safeguarding the integrity of pharmaceutical distribution plays a crucial role in preventing the circulation of counterfeit drugs. Traditional drug supply chain systems have no effective mechanism to ensure the traceability and integrity of the drug supply chain process. Various ICT-based solutions have been adopted previously and numerous regulations were passed by drug regulatory authorities to ensure interoperable and traceable drug distribution process. However, these solutions could not guarantee a tamper-proof history of supply chain process. Blockchain technology has gained traction in many diverse fields including real-estate, identity management, digital voting, healthcare, etc. Therefore, we have proposed a decentralized framework to insure the authenticity and traceability of the drugs. Moreover, this study presents an overview of how blockchain technology can be employed to counter the fake medicines circulation by embedding the traceability in drug supply chain process. Comparison of various blockchain-based solutions and open research challenges in this regard are also presented.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Security Standards and use of Cyberinfrastructure Systems in Clinical Research 23

Security Standards and use of Cyberinfrastructure Systems in...

引用

2023 Practice and Experience in Advanced Research computing, PEARC 2023

作者： Shankar, Anurag Hancock, David Michael, Scott Huber, Laura Butler, Joseph Henschel, Robert Stewart, Craig Center for Applied Cybersecurity Research Indiana University BloomingtonIN United States Research Technologies Indiana University BloomingtonIN United States Department of Computer Science Luddy School of Informatics Computing and Engineering Indiana University BloomingtonIN United States

ISBN: (纸本)9781450399852

Many areas of biomedical research, including precision health research, inherently involve the use of IT systems properly secured to store and analyze protected health information (PHI). Managing cybersecurity policies in ways that allow such storage and analysis of PHI on computing systems is involved, but so doing can facilitate important research. In 2005, Indiana University established the ability to analyze PHI (HIPAA alignment) as the minimal security level for its High Performance computing (HPC) systems and research storage systems. Starting in 2010, REDCap was also offered as a central service, providing capabilities for conducting surveys that collect PHI and storing those data in ways that align with HIPAA. Usage data shows that providing computational and data resources aligned with HIPAA has increased the use of advanced research cyberinfrastructure over time. We explain some of our tactics in providing secure environments for handling protected health information and offer suggestions that other cyberinfrastructure centers may find useful in setting strategies for serving medical researchers. © 2023 ACM.

关键词： Clinical research

来源：评论

学校读者我要写书评

暂无评论

AOCC-FL: Federated Learning with Aligned Overlapping via Calibrated Compensation 42

AOCC-FL: Federated Learning with Aligned Overlapping via Cal...

引用

42nd IEEE International Conference on computer Communications, INFOCOM 2023

作者： Wang, Haozhao Xu, Wenchao Fan, Yunfeng Li, Ruixuan Zhou, Pan Huazhong University of Science and Technology School of Computer Science and Technology Wuhan China The Hong Kong Polytechnic University Department of Computing Hong Kong Huazhong University of Science and Technology Hubei Key Laboratory of Distributed System Security Hubei Engineering Research Center on Big Data Security School of Cyber Science and Engineering Wuhan China

ISBN: (纸本)9798350334142

Federated Learning enables collaboratively model training among a number of distributed devices with the coordination of a centralized server, where each device alternatively performs local gradient computation and communication to the server. FL suffers from significant performance degradation due to the excessive communication delay between the server and devices, especially when the network bandwidth of these devices is limited, which is common in edge environments. Existing methods overlap the gradient computation and communication to hide the communication latency to accelerate the FL training. However, the overlapping can also lead to an inevitable gap between the local model in each device and the global model in the server that seriously restricts the convergence rate of learning process. To address this problem, we propose a new overlapping method for FL, AOCC-FL, which aligns the local model with the global model via calibrated compensation such that the communication delay can be hidden without deteriorating the convergence performance. Theoretically, we prove that AOCC-FL admits the same convergence rate as the non-overlapping method. On both simulated and testbed experiments, we show that AOCC-FL achieves a comparable convergence rate relative to the non-overlapping method while outperforming the state-of-the-art overlapping methods. © 2023 IEEE.

关键词： Learning systems

来源：评论

学校读者我要写书评

暂无评论

Cloud-Based Mixed Reality System in Machinery Vocational School: ICT Blueprint

Cloud-Based Mixed Reality System in Machinery Vocational Sch...

引用

2022 International Conference on science and Technology, ICOSTECH 2022

作者： Suryodiningrat, Satrio Pradono Gaol, Ford L. Supangkat, Suhono Harso Ranti, Benny Bina Nusantara University Binus Graduate Program Doctor of Computer Science Computer Science Department Jakarta Indonesia Bina Nusantara University Faculty of Computing and Media Computer Science Department Jakarta Indonesia Bina Nusantara University Jakarta Doctor of Computer Science Graduate Program Jakarta Indonesia Smart City and Community Innovation Center Bandung Institute of Technology Bandung Indonesia

ISBN: (纸本)9781665479851

Education is always the most important part of our lives. To make education make a significant impact on today's learners, teaching and learning must evolve to match with today's learner's profile. Any education including vocational education has an important role in economic development and one of the ways to maximize the contribution of vocational educations is strengthening the link and match with the world of work [1]. As the technology also keeps evolving and it is getting more and more advanced, one technology that has significant traction lately is the mixed reality system and is often used for education purposes. Most big ICT companies like Microsoft, and HTC are racing to develop mixed reality systems to help the teaching and learning process to be more effective and efficient. Where everything is cloud-based, having a cloud-based mixed reality system is just makes sense but will have a great challenge and considered as a big ICT project. One of the ways to make the ICT project successful is to make a blueprint before it is started. The objective of this paper is how to develop an ICT Blueprint for implementing a cloud-based mixed reality system for machinery vocational schools to make the project have a higher success rate. © 2022 IEEE.

关键词： Blueprints

来源：评论

学校读者我要写书评

暂无评论

Accumulative Time Based Ranking Method to Reputation Evaluation in Information Networks

引用

Journal of computer science & Technology 2022年第4期37卷 960-974页

作者： Hao Liao Qi-Xin Liu Ze-Cheng Huang Ke-Zhong Lu Chi Ho Yeung Yi-Cheng Zhang National Engineering Laboratory on Big Data System Computing Technology College of Computer Science and SoftwareEngineeringShenzhen UniversityShenzhen 518060China Guangdong Province Key Laboratory of Popular High Performance Computers Shenzhen UniversityShenzhen 518060China Guangdong Province Engineering Center of China-Made High Performance Data Computing System College of ComputerScience and Software EngineeringShenzhen UniversityShenzhen 518060China Institute of Big Data Intelligent Management and Decision Shenzhen UniversityShenzhen 518060China Department of Science and Environmental Studies The Education University of Hong KongHong Kong 999077China Department of Physics University of FribourgFribourg 1700Switzerland

Due to over-abundant information on the Web, information filtering becomes a key task for online users to obtain relevant suggestions and how to extract the most related item is always a key topic for researchers in various fields. In this paper, we adopt tools used to analyze complex networks to evaluate user reputation and item quality. In our proposed Accumulative Time Based Ranking (ATR) algorithm, we take into account the growth record of the network to identify the evolution of the reputation of users and the quality of items, by incorporating two behavior weighting factors which can capture the hidden facts on reputation and quality dynamics for each user and item respectively. Our proposed ATR algorithm mainly combines the iterative approach to rank user reputation and item quality with temporal dependence compared with other reputation evaluation methods. We show that our algorithm outperforms other benchmark ranking algorithms in terms of precision and robustness on empirical datasets from various online retailers and the citation datasets among research publications. Therefore, our proposed method has the capability to effectively evaluate user reputation and item quality.

关键词： temporal network behavior dynamics reputation evaluation ranking algorithm

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：