检索结果-内蒙古大学图书馆

Optimizing Reservoir Computing Based on an Alternating Input-Driven Spin-Torque Oscillator

Physical Review Applied 2023年第2期20卷 024069-024069页

作者： Xuezhao Wu Zihan Tong Qiming Shao Department of Electronic and Computer Engineering The Hong Kong University of Science and Technology Clear Water Bay Hong Kong SAR China State Key Laboratory of Advanced Displays and Optoelectronics Technologies The Hong Kong University of Science and Technology Clear Water Bay Hong Kong SAR China ACCESS – AI Chip Center for Emerging Smart Systems InnoHK Centers Hong Kong Science Park Hong Kong SAR China Department of Physics and IAS Center for Quantum Technologies The Hong Kong University of Science and Technology Hong Kong SAR China Guangdong-Hong Kong-Macao Joint Laboratory for Intelligent Micro-Nano Optoelectronic Technology The Hong Kong University of Science and Technology Hong Kong SAR China

Neurons in the brain show nonlinear oscillatorlike behaviors, inspiring the development of neuromorphic computing with spin-torque oscillators for their low power consumption and high integration density. The time-multiplexing technique enables utilizing only one device to achieve neuromorphic computing, further reducing power consumption and simplifying device fabrication. However, optimizing such systems remains a challenge and needs further development. A controversial hypothesis claims that dynamic systems possess the optimal computational capacity to address the information at the dynamical phase transition. Here, we apply this hypothesis to improve a reservoir computing system based on an alternating input-driven spin-torque oscillator. The input-driven property of reservoir computing allows us to exploit the dynamics of a spin-torque oscillator by manipulating the input stream. By tuning the ac amplitude in inputs, the system can exhibit diversified dynamic regimes indicated by the largest Lyapunov exponent and echo-state property spectra. Our findings demonstrate that the proper configuration of ac inputs for reservoir computing based on a single spin-torque oscillator can adjust its information-processing capacity profile and enhance its computational performance for specific tasks.

关键词： Magnetization dynamics Network phase transitions Neuromorphic computing Spintronics Coupled oscillators Devices for digital logic, storage & processing Spin torque nano-oscillators Chaos & nonlinear dynamics

来源：评论

学校读者我要写书评

暂无评论

Mixture of Experts for Network Optimization: A Large Language Model-enabled Approach

arXiv

引用

arXiv 2024年

作者： Du, Hongyang Liu, Guangyuan Lin, Yijing Niyato, Dusit Kang, Jiawen Xiong, Zehui Kim, Dong In The School of Computer Science and Engineering Nanyang Technological University Singapore639798 Singapore The State Key Laboratory of Networking and Switching Technology Beijing University of Posts and Telecommunications Beijing100876 China The School of Automation Guangdong University of Technology Key Laboratory of Intelligent Information Processing and System Integration of IoT Ministry of Education Guangzhou510006 China Guangdong-HongKong-Macao Joint Laboratory for Smart Discrete Manufacturing Guangzhou510006 China The Pillar of Information Systems Technology and Design Singapore University of Technology and Design Singapore487372 Singapore The Department of Electrical and Computer Engineering Sungkyunkwan University Suwon16419 Korea Republic of

Optimizing various wireless user tasks poses a significant challenge for networking systems because of the expanding range of user requirements. Despite advancements in Deep Reinforcement Learning (DRL), the need for customized optimization tasks for individual users complicates developing and applying numerous DRL models, leading to substantial computation resource and energy consumption and can lead to inconsistent outcomes. To address this issue, we propose a novel approach utilizing a Mixture of Experts (MoE) framework, augmented with Large Language Models (LLMs), to analyze user objectives and constraints effectively, select specialized DRL experts, and weigh each decision from the participating experts. Specifically, we develop a gate network to oversee the expert models, allowing a collective of experts to tackle a wide array of new tasks. Furthermore, we innovatively substitute the traditional gate network with an LLM, leveraging its advanced reasoning capabilities to manage expert model selection for joint decisions. Our proposed method reduces the need to train new DRL models for each unique optimization problem, decreasing energy consumption and AI model implementation costs. The LLM-enabled MoE approach is validated through a general maze navigation task and a specific network service provider utility maximization task, demonstrating its effectiveness and practical applicability in optimizing complex networking systems. © 2024, CC BY.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Adaptive Discriminative Regularization for Visual Classification

arXiv

引用

arXiv 2022年

作者： Zhao, Qingsong Wang, Yi Dou, Shuguang Gong, Chen Wang, Yin Zhao, Cairong Ji, Xiaopeng Jiang, Xinyang Department of Computer Science and Technology Tongji University Shanghai201804 China Shanghai AI Laboratory Shanghai200232 China Key Laboratory of Intelligent Perception and Systems for High-Dimensional Information School of Computer Science and Engineering Ministry of Education Nanjing University of Science and Technology Nanjing210094 China The State Key Lab of CAD&CG Zhejiang University China Shanghai China

How to improve discriminative feature learning is central in classification. Existing works address this problem by explicitly increasing inter-class separability and intra-class compactness, whether by constructing positive and negative pairs for contrastive learning or posing tighter class separating margins. These methods do not exploit the similarity between different classes as they adhere to independent identical distributions assumption in data. In this paper, we embrace the real-world data distribution setting that some classes share semantic overlaps due to their similar appearances or concepts. Regarding this hypothesis, we propose a novel regularization to improve discriminative learning. We first calibrate the estimated highest likelihood of one sample based on its semantically neighboring classes, then encourage the overall likelihood predictions to be deterministic by imposing an adaptive exponential penalty. As the gradient of the proposed method is roughly proportional to the uncertainty of the predicted likelihoods, we name it adaptive discriminative regularization (ADR), trained along with a standard cross entropy loss in classification. Extensive experiments demonstrate that it can yield consistent and non-trivial performance improvements in a variety of visual classification tasks (over 10 benchmarks). Furthermore, we find it is robust to long-tailed and noisy label data distribution. Its flexible design enables its compatibility with mainstream classification architectures and losses. © 2022, CC BY.

关键词： Image classification

来源：评论

学校读者我要写书评

暂无评论

Online Attentive Kernel-Based Temporal Difference Learning

arXiv

引用

arXiv 2022年

作者： Yang, Guang Chen, Xingguo Yang, Shangdong Wang, Huihui Dong, Shaokang Gao, Yang The the Jiangsu Key Laboratory of Big Data Security & Intelligent Processing Nanjing University of Posts and Telecommunications National Engineering Laboratory for Agri-Product Quality Traceability Beijing Technology and Business University China The State Key Laboratory for Novel Software Technology Nanjing University China The PCA Lab Key Lab of Intelligent Perception and Systems for High-Dimensional Information Ministry of Education Jiangsu Key Lab of Image and Video Understanding for Social Security School of Computer Science and Engineering Nanjing University of Science and Technology China

With rising uncertainty in the real world, online Reinforcement Learning (RL) has been receiving increasing attention due to its fast learning capability and improving data efficiency. However, online RL often suffers from complex Value Function Approximation (VFA) and catastrophic interference, creating difficulty for the deep neural network to be applied to an online RL algorithm in a fully online setting. Therefore, a simpler and more adaptive approach is introduced to evaluate value function with the kernel-based model. Sparse representations are superior at handling interference, indicating that competitive sparse representations should be learnable, non-prior, non-truncated and explicit when compared with current sparse representation methods. Moreover, in learning sparse representations, attention mechanisms are utilized to represent the degree of sparsification, and a smooth attentive function is introduced into the kernel-based VFA. In this paper, we propose an Online Attentive Kernel-Based Temporal Difference (OAKTD) algorithm using two-timescale optimization and provide convergence analysis of our proposed algorithm. Experimental evaluations showed that OAKTD outperformed several Online Kernel-based Temporal Difference (OKTD) learning algorithms in addition to the Temporal Difference (TD) learning algorithm with Tile Coding on public Mountain Car, Acrobot, CartPole and Puddle World tasks. Copyright © 2022, The Authors. All rights reserved.

关键词： Learning algorithms

来源：评论

学校读者我要写书评

暂无评论

A Time-Series Higher-Order Network Hierarchical Evolution Model for Dynamic Interaction of Nodes

SSRN

引用

SSRN 2022年

作者： Yang, Xiaoteng Wu, Zhenqiang Zhang, WenBo Javaid, Shumaila School of Computer Science Shaanxi Normal University Xi’an710119 China Key Laboratory of Modern Teaching Technology Ministry of Education Xi’an710062 China Department of Control Science and Engineering College of Electronics and Information Engineering Tongji University Shanghai201804 China Frontiers Science Center for Intelligent Autonomous Systems Shanghai201210 China

The evolution of hierarchies in complex systems is an exciting phenomenon since hierarchies exist in many social and biological networks. So far, many approaches have been proposed, however, they are either based on the locality of the nodes or the cluster relationships between networks. These approaches often use traditional time-series network modeling methods that focus only on local static relationships between different nodes but ignore the dynamic interactions between real network nodes. Local static and dynamic global relationships should be considered to understand better the generation and maintenance of hierarchical phenomena in networks. To address this problem, we propose a long-term, flexible hierarchical evolution model for explaining the hierarchical phenomena arising in networks. Our model can integrate both local static and dynamic global information in the network and use time and higher-order inter-node interactions to integrate static and dynamic relationships. New interactions in the network are formed based on preference information between nodes, while old links fade out over time. In addition, we derive critical points for the hierarchical evolution of the system over long periods. Importantly, our model allows the comparison of network rank formation across different data as a way to perform rank evolution studies. © 2022, The Authors. All rights reserved.

关键词： Time series

来源：评论

学校读者我要写书评

暂无评论

Dfier: A Directed Vulnerability Verifier for Ethereum Smart Contracts

SSRN

引用

SSRN 2023年

作者： Wang, Zeli Dai, Weiqi Li, Ming Choo, Kim-Kwang Raymond Zou, Deqing Chongqing Key Laboratory of Computational Intelligence Key Laboratory of Big Data Intelligent Computing Key Laboratory of Cyberspace Big Data Intelligent Security Ministry of Education College of Computer Science and Technology Chongqing University of Posts and Telecommunications Chongqing40065 China National Engineering Research Center for Big Data Technology and System Services Computing Technology and System Lab Hubei Engineering Research Center on Big Data Security Hubei Key Laboratory of Distributed System Security Hubei Engineering Research Center on Big Data Security School of Cyber Science and Engineering Huazhong University of Science and Technology Wuhan430074 China Department of Information Systems and Cyber Security University of Texas at San Antonio San Antonio United States

Existing smart contract vulnerability identification approaches mainly focus on complete program detection. Consequently, lots of known potentially vulnerable locations need manual verification, which is energy-exhaustive and error-prone. Fuzzing is an effective way to detect vulnerabilities. However, current fuzzers cannot effectively probe these suspicious targets because they overlook two key factors: exploring all paths toward the targets and providing high-quality directed seed inputs. This paper presents a Directed vulnerability veriFier (DFier), elaborating effective transaction sequences with directed inputs for the fuzzer, to focus on analyzing target paths and automatically detect whether the given locations are vulnerable. Specifically, DFier fuses the effectiveness of static analysis and the efficiency of fuzzing. The static analysis helps locate target paths and generate directed inputs for the fuzzer under three heuristic strategies. The first prefers inputs with the larger possibility of triggering potential vulnerabilities based on target features. The second guides the value mutation to satisfy the path constraints. The third adequately leverages execution information to help find exploits faster. Extensive experiments demonstrate that DFier is effective in verifying contract security, compared with three existing contract fuzzers (i.e., contractFuzzer, sFuzz, and conFuzzius). And the performance losses are in an acceptable range. © 2023, The Authors. All rights reserved.

关键词： Smart contract

来源：评论

学校读者我要写书评

暂无评论

Utilization of machine learning and neural networks to optimize the enclosure angle, magnetic field, and radiation parameter for mixed convection of hybrid nanofluid flow next to assess environmental impact

引用

engineering Analysis with Boundary Elements 2023年 146卷 252-262页

作者： Tao Hai Hayder A. Dhahad Masood Ashraf Ali Vishal Goyal Sattam Fahad Almojil Abdulaziz Ibrahim Almohana Abdulrhman Fahmi Alali Khaled Twfiq Almoalimi Farah Qasim Ahmed Alyousuf School of Computer and Information Qiannan Normal University for Nationalities Duyun Guizhou 558000 China Key Laboratory of Complex Systems and Intelligent Optimization of Guizhou Duyun Guizhou 558000 China Institute for Big Data Analytics and Artificial Intelligence (IBDAAI) Universiti Teknologi MARA Shah Alam Selangor 40450 Malaysia Mechanical Engineering Department University of Technology Baghdad Iraq Department of Industrial Engineering College of Engineering Prince Sattam bin Abdulaziz University Alkharj 16273 Saudi Arabia Department of Electronics and Communication Engineering GLA University Mathura India Department of Civil Engineering College of Engineering King Saud University P.O. Box 800 Riyadh 11421 Saudi Arabia Department of Information Technology College of Engineering and Computer Science Lebanese French University Kurdistan Region Iraq

This paper examines the mixed convective heat transfer (HTR) of nanofluid (NFD) flow in a rectangular enclosure with the upper moving wall numerically. The lower wall has a high temperature and a number of semi-circular obstacles with the same temperature are installed on it. The upper moving wall has a low temperature and the other two walls are insulated. The enclosure can change from horizontal to vertical. Radiation HTR is considered in the enclosure and there is a magnetic field (MGF) that can change the angle from horizontal to vertical affecting the NFD. This study is carried out for different angles of the enclosure and MGF from horizontal to vertical for radiation parameters (RDP) of 0 to 3 and a constant MGF with Hartmann number of 20 and Richardson number of 10. The aim is to estimate the Nusselt number (Nu), entropy generation (ETG), and Bejan number (Be). The SIMPLE algorithm is utilized using FORTRAN software, and optimization is done using artificial intelligence to find the maximum and minimum output values. The results demonstrate that the maximum value of Nu and Bes corresponds to the MGF angle and enclosure angle of 90°. The minimum value of the Nu and the maximum amount of ETG corresponds to the horizontal MGF and horizontal enclosure when the RDP is 1.5. An increment in the RDP enhances the amount of Nu. The maximum amount of ETG, i.e. 12.87, corresponds to the enclosure with an angle of 45° for the horizontal MGF and the absence of RDP. corresponds to the enclosure with an angle of 45° for the horizontal MGF and the absence of RDP. It was also found that most environmental impacts, and hence values for different environmental factors, arise from the production of nanoparticles; thus, it is a significant contributor to environmental impacts.

关键词： Machine learning Entropy generation Numerical calculations Convection Nanofluid

来源：评论

学校读者我要写书评

暂无评论

Spatial-Temporal Enhancement Based Tourism Demand Forecasting: A Deep Model for a Case in Urban Area

SSRN

引用

SSRN 2022年

作者： Dong, Yunxuan Zhou, Binggui Yang, Guanghua Hou, Fen Hu, Zheng Ma, Shaodan School of Intelligent Systems Science and Engineering Jinan University Zhuhai519070 China State Key Laboratory of Internet of Things for Smart City the Department of Electrical and Computer Engineering University of Macau 999078 China State Key Laboratory of Network and Switching Technology Beijing University of Posts and Telecommunications Beijing100876 China

Accurate tourism demand forecasting requires the learning of the spatial-temporal features of tourism demand, which is challenging because the features may constantly change owing to human behavior. In this study, we developed a spatial-temporal feature enhancement model to ensure the integrity of the features of tourism demand. By modeling the tourism system as an undirected graph, we employ the steady-state analysis method of the undirected graph to extract the spatial-temporal features of the tourism system, and the extracted features are employed to construct the enhancement feature vector. However, the spatial-temporal features in the enhancement feature vector are often sparse. We propose a deep model with convolutional filters to ensure the feature learning ability for sparse features;however, there are challenges to using the high-dimensional convolutional filters to learn the features of an one-dimensional time series. Therefore, we convert the feature series into an image series while maintaining the relationship of the spatial-temporal features. The digital footprints of tourists from the urban area of Zhuhai were used to verify the effectiveness of the method. Numerical experiments show that the proposed model can effectively capture the evolution process of the tourism demand system, and the forecasting performance is better than that of state-of-the-art tourism demand forecasting models. © 2022, 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

Reducing Idleness in Financial Cloud Services via Multi-objective Evolutionary Reinforcement Learning based Load Balancer

arXiv

引用

arXiv 2023年

作者： Yang, Peng Zhang, Laoming Liu, Haifeng Li, Guiying Department of Statistics and Data Science Southern University of Science and Technology Shenzhen518055 China Guangdong Provincial Key Laboratory of Brain-Inspired Intelligent Computation Department of Computer Science and Engineering Southern University of Science and Technology Shenzhen518055 China Academy for Advanced Interdisciplinary Studies Southern University of Science and Technology Shenzhen518055 China Guangdong OPPO Mobile Telecommunications Corp. Ltd Shenzhen518052 China Research Institute of Trustworthy Autonomous Systems Southern University of Science and Technology Shenzhen518055 China

In recent years, various companies have started to shift their data services from traditional data centers to the cloud. One of the major motivations is to save on operational costs with the aid of cloud elasticity. This paper discusses an emerging need from financial services to reduce the incidence of idle servers retaining very few user connections, without disconnecting them from the server side. This paper considers this need as a bi-objective online load balancing problem. A neural network based scalable policy is designed to route user requests to varied numbers of servers for the required elasticity. An evolutionary multi-objective training framework is proposed to optimize the weights of the policy. Not only is the new objective of idleness is reduced by over 130% more than traditional industrial solutions, but the original load balancing objective itself is also slightly improved. Extensive simulations with both synthetic and real-world data help reveal the detailed applicability of the proposed method to the emergent problem of reducing idleness in financial services. Copyright © 2023, The Authors. All rights reserved.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Rank-Based Filter Pruning for Real-Time Deep Uav Trackers

SSRN

引用

SSRN 2022年

作者： Wang, Xucheng Zeng, Dan Zhao, Qijun Li, Shuiwang Guangxi Key Laboratory of Embedded Technology and Intelligent System Guilin University of Technology Guilin541001 China Research Institute of Trustworthy Autonomous Systems Department of Computer Science and Engineering Southern University of Science and Technology Shenzhen518000 China Sichuan University Chengdu610000 China

UAV tracking is an emerging task and has wide potential applications in such as agriculture, navigation, entertainment and public security. However, the limitations of computing resources, battery capacity, and maximum load of UAV hinder the deployment of DL-based tracking algorithms on UAV. In contrast to deep learning trackers, discriminative correlation filters (DCF)-based trackers stand out in the UAV tracking community because of their high efficiency. However, their precision is usually much lower than trackers based on deep learning. Model compression is a promising way to narrow the gap (i.e., effciency, precision) between DCF- and deep learning- based trackers, which has not caught much attention in UAV tracking community. In this paper, We propose the P-SiamFC++ tracker, which is the first to use rank-based filter pruning to compress the SiamFC++ model, achieving a remarkable balance between efficiency and precision. Our method is general and may encourage further studies on UAV tracking with model compression in the future. Extensive experiments on four UAV benchmarks, including UAV123@10fps, DTB70, UAVDT and Vistrone2018, show that P-SiamFC++ tracker significantly outperforms state-of-the-art UAV tracking methods. © 2022, The Authors. All rights reserved.

关键词： Unmanned aerial vehicles (UAV)

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：