检索结果-内蒙古大学图书馆

Synthesizing Decentralized Controllers With Graph neural networks and Imitation Learning

IEEE TRANSACTIONS ON SIGNAL processing 2022年 70卷 1932-1946页

作者： Gama, Fernando Li, Qingbiao Tolstaya, Ekaterina Prorok, Amanda Ribeiro, Alejandro Rice Univ Dept Comp & Elect Engn Houston TX 77005 USA Univ Penn Dept Elect & Syst Engn Philadelphia PA 19104 USA Univ Cambridge Dept Comp Sci & Technol Cambridge CB2 1TN England

Dynamical systems consisting of a set of autonomous agents face the challenge of having to accomplish a global task, relying only on local information. While centralized controllers are readily available, they face limitations in terms of scalability and implementation, as they do not respect the distributed information structure imposed by the network system of agents. Given the difficulties in finding optimal decentralized controllers, we propose a novel framework using graph neural networks (GNNs) to learn these controllers. GNNs are well-suited for the task since they are naturally distributed architectures and exhibit good scalability and transferability properties. We show that GNNs learn appropriate decentralized controllers by means of imitation learning, leverage their permutation invariance properties to successfully scale to larger teams and transfer to unseen scenarios at deployment time. The problems of flocking and multi-agent path planning are explored to illustrate the potential of GNNs in learning decentralized controllers.

关键词： Decentralized control Scalability Robot sensing systems Graph neural networks Training Costs Trajectory Decentralized control graph neural networks graph signal processing flocking path planning

来源：评论

学校读者我要写书评

暂无评论

Recurrent Convolution neural network with Group-Wise Enhancement for Student English Classroom Learning Behavior Recognition

Recurrent Convolution Neural Network with Group-Wise Enhance...

引用

2024 International Conference on distributed Systems, Computer networks and Cybersecurity, ICDSCNC 2024

作者： Lang, Shanshan Sui, Xin Changchun Humanities and Sciences College School of Foreign Language and Literature Changchun China Changchun Humanities and Sciences College Public Computer Department Changchun China

ISBN: (纸本)9798350375442

Due to the increase in the number of students joining colleges across the world, current college classrooms are larger, and this has a stressed university lecturer teaching those large classes and conducting lectures efficiently. Therefore, the Recurrent Convolution neural network with Group-Wise Enhancement (RCNNGWE) is proposed in this research for English classroom learning behavior recognition. The bilateral filter is used for preprocessing to reduce noise when reserving edges and enhancing the visual feature clarity. Then, the feature extraction is performed by Scale-Invariant Feature Transform (SIFT) which extracts visual features to scale and orientation. The RCNNGWE works well because it is able to simulate environment within the classroom as well as the different behaviors of students in the class. To further propagate key behavioral patterns the proposed model, need to localize groupwise operations which helps to improve recognition accuracy. The GWE continues to enrich the functionality of the proposed model by grouping inputs and processing them in a separate aggregate, which enables identification of behaviors of students. The RCNNGWE obtained 93.75% accuracy, 87.27% f1-score, 86.41% precision, 87.26% mean Average Precision (mAP) and 91.64% recall for Student Learning Behavior (SLB) dataset. © 2024 IEEE.

关键词： bilateral filter english classroom learning behavior group-wise enhancement recurrent convolution neural network scale-invariant feature transform

来源：评论

学校读者我要写书评

暂无评论

Physics-constrained adaptive kernel interpolation for region-to-region acoustic transfer function: a Bayesian approach

引用

EURASIP JOURNAL ON AUDIO SPEECH AND MUSIC processing 2024年第1期2024卷 43页

作者： Ribeiro, Juliano G. C. Koyama, Shoichi Saruwatari, Hiroshi Univ Tokyo Grad Sch Informat Sci & Technol 7-3-1 HongoBunkyo Ku Tokyo 1138654 Japan Natl Inst Informat Digital Content & Media Sci Res Div 2-1-2 HitotsubashiChiyoda Ku Tokyo 1018430 Japan

A kernel interpolation method for the acoustic transfer function (ATF) between regions constrained by the physics of sound while being adaptive to the data is proposed. Most ATF interpolation methods aim to model the ATF for fixed source by using techniques that fit the estimation to the measurements while not taking the physics of the problem into consideration. We aim to interpolate the ATF for a region-to-region estimation, meaning we account for variation of both source and receiver positions. By using a very general formulation for the reproducing kernel function, we have created a kernel function that considers both directed and residual fields as two separate kernel functions. The directed field kernel considers a sparse selection of reflective field components with large amplitudes and is formulated as a combination of directional kernels. The residual field is composed of the remaining densely distributed components with lower amplitudes. Its kernel weight is represented by a universal approximator, a neural network, in order to learn patterns from the data freely. These kernel parameters are learned using Bayesian inference both under the assumption of Gaussian priors and by using a Markov chain Monte Carlo simulation method to perform inference in a more directed manner. We compare all established kernel formulations with each other in numerical simulations, showing that the proposed kernel model is capable of properly representing the complexities of the ATF.

关键词： Acoustic transfer function Region-to-region interpolation Kernel methods neural networks Bayesian inference

来源：评论

学校读者我要写书评

暂无评论

Cluster-based downscaling of precipitation using Kolmogorov-Arnold neural networks and CMIP6 models: Insights from Oman

引用

JOURNAL OF ENVIRONMENTAL MANAGEMENT 2025年 380卷 124971页

作者： Mardy, Ali Nikoo, Mohammad Reza Zamani, Mohammad G. Al-Rawas, Ghazi Nazari, Rouzbeh Simunek, Jiri Sana, Ahmad Gandomi, Amir H. KNToosi Univ Fac Civil Water & Environm Engn Dept Water Resources Engn Tehran Iran Sultan Qaboos Univ Dept Civil & Architectural Engn Muscat Oman Univ Memphis Herff Coll Engn Dept Civil Engn Memphis TN 38152 USA Univ Calif Riverside Dept Environm Sci Riverside CA USA PNG Univ Technol Sch Civil Engn Lae Papua N Guinea Univ Technol Sydney Dept Engn & IT Ultimo NSW 2007 Australia Obuda Univ Univ Res & Innovat Ctr EKIK H-1034 Budapest Hungary

Accurate precipitation predictions are crucial for addressing climate change impacts on water resources, especially in arid regions like Oman. Therefore, this study evaluates three machine learning models-Random Forest (RF), Multilayer Perceptron neural networks (MLP-ANN), and Kolmogorov-Arnold neural networks (KANNs)- to downscale and predict precipitation patterns under climate scenarios SSP1-2.6, SSP2-4.5, and SSP5-8.5. We assessed each model's ability to reproduce past trends and predict future precipitation using historical data from 1995 to 2014 and projections from 2020 to 2099. The KANN model demonstrated exceptional proficiency in forecasting extreme precipitation occurrences, especially in the most severe scenario (SSP5-8.5). The MLP-ANN model offered a balanced methodology, yielding dependable forecasts that are adaptive to fluctuating situations, even amongst small increases in precipitation and uncertainty. The RF model generated the most reliable forecasts, suggesting small increases in future precipitation while closely correlating with historical data. The study indicates distinct seasonal patterns, with peak precipitation occurring during the monsoon season from June to August. The RF model predicted more intense and uniformly distributed precipitation during this period, demonstrating its advanced data processing capabilities. The geographical patterns predicted by each model exhibited temporal stability, highlighting their consistent reliability, which is essential for precise climate predictions. This comparative research highlights the significance of choosing a suitable machine learning model according to distinct forecasting requirements. Effective downscaling methodologies are essential for informed water resources management, particularly in areas susceptible to cyclones and water shortages. These results provide essential direction for policymakers to improve climate resilience, optimize water infrastructure, and formulate adaptation measur

关键词： Statistical downscaling Machine learning Precipitation Clustering Kolmogorov-Arnold neural network

来源：评论

学校读者我要写书评

暂无评论

Why Go Full? Elevating Federated Learning Through Partial network Updates 38

Why Go Full? Elevating Federated Learning Through Partial Ne...

引用

38th Conference on neural Information processing Systems, NeurIPS 2024

作者： Wang, Haolin Liu, Xuefeng Niu, Jianwei Guo, Wenkai Tang, Shaojie State Key Laboratory of Virtual Reality Technology and Systems School of Computer Science and Engineering Beihang University Beijing China Center for AI Business Innovation Department of Management Science and Systems University at Buffalo BuffaloNY United States Zhongguancun Laboratory Beijing China

Federated learning is a distributed machine learning paradigm designed to protect user data privacy, which has been successfully implemented across various scenarios. In traditional federated learning, the entire parameter set of local models is updated and averaged in each training round. Although this full network update method maximizes knowledge acquisition and sharing for each model layer, it prevents the layers of the global model from cooperating effectively to complete the tasks of each client, a challenge we refer to as layer mismatch. This mismatch problem recurs after every parameter averaging, consequently slowing down model convergence and degrading overall performance. To address the layer mismatch issue, we introduce the FedPart method, which restricts model updates to either a single layer or a few layers during each communication round. Furthermore, to maintain the efficiency of knowledge acquisition and sharing, we develop several strategies to select trainable layers in each round, including sequential updating and multi-round cycle training. Our theoretical analysis and experimental results show that the FedPart method significantly outperforms traditional full-network update strategies, achieving faster convergence, greater accuracy, and reduced communication and computational overhead. © 2024 neural information processing systems foundation. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Research on Computer network Information System Security Model and Mathematical Evaluation 2

Research on Computer Network Information System Security Mod...

引用

2nd International Conference on Mechatronics, IoT and Industrial Informatics, ICMIII 2024

作者： Liu, Yanwu Beijing Polytechnic Beijing China

ISBN: (纸本)9798350386639

This paper firstly proposes a distributed network Information filtering architecture based on multi-level, multi-strategy and scalability. The architecture of this model includes the functions of analyzing, filtering and forwarding the Internet information data, improving the data forwarding rate and reducing the data transmission delay under the premise of ensuring the security of the information data, and the filter program can be installed in a plug-in mode, with good openness and scalability. This architecture can modular combination and configuration of each function according to specific requirements, so that they can cooperate in a decentralized environment, thus greatly improving the efficiency of the entire system. Combining the fuzzy mathematics theory with BP neural network, the method of risk assessment in information system is established. This method is a combination of qualitative and quantitative methods, through the fuzzy processing of the risks stored in the information system and then using BP neural network to train. This method can improve the efficiency and accuracy of risk assessment of information system. © 2024 IEEE.

关键词： Information filtering

来源：评论

学校读者我要写书评

暂无评论

Secure sharing of industrial IoT data based on distributed trust management and trusted execution environments: a federated learning approach

引用

neural COMPUTING & APPLICATIONS 2023年第29期35卷 21499-21509页

作者： Zheng, Wei Cao, Yang Tan, Haining Beijing Univ Technol Fac Informat Technol Beijing Peoples R China Beijing Wuzi Univ Sch Informat Beijing Peoples R China Shenzhen High Tech Ind Pk Informat Network Co Ltd Shenzhen Peoples R China

Industrial Internet of Things (I-IoT) has become an emerging driver to operate industrial systems and a primary empowerer to future industries. With the advanced technologies such as artificial intelligence (AI) and machine learning widely used in IoT, the Industrial IoT is also witnessing changes driven by new technologies. Generally, AI technologies require centralized data collection and processing to learn from the data to obtain viable models for application. In industrial IoT, data security and privacy problems associated with reliable and interconnected end devices are being faced and reliable solutions are urgently needed. A trusted execution environment in IoT devices is gradually becoming a feasible approach, and a distributed solution is a natural choice for artificial intelligence technologies in I-IoT. Moreover, Federated Learning as a distributed machine learning paradigm with privacy-preserving properties can be used in I-IoT. This paper introduces a feasible secure data circulation and sharing scheme for I-IoT devices in a trusted implementation platform by employing federated learning. The suggested framework has proved to be efficient, reliable, and accurate.

关键词： Industrial internet of things Trusted execution environment Federated learning Data security sharing

来源：评论

学校读者我要写书评

暂无评论

In-network Convolution in Grid Shaped Sensor networks

引用

JOURNAL OF WEB ENGINEERING 2022年第1期21卷 75-96页

作者： Hrovatin, Niki Tosic, Aleksandar Vicic, Jernej Univ Primorska Fac Math Nat Sci & Informat Technol Koper Slovenia InnoRenew CoE Izola Slovenia Slovenian Acad Sci & Arts Res Ctr Fran Ramovs Inst Ljubljana Slovenia

Gathering information is the primary purpose of a Sensor network. The task is performed by spatially distributed nodes equipped with sensing, processing, and communication capabilities. However, data gathered from a sensor network must be processed, and often the collective computation capability of nodes forming the sensor network is neglected in favor of data processing on cloud systems. Nowadays, Edge Computing has emerged as a new paradigm aiming to migrate data processing close to data sources. In this contribution, we focus on the development of a sensor network designed to detect a person's fall. We named this sensor network the smart floor. Fall detection is tackled with a Convolutional neural network, and we propose an approach for in-network processing of convolution layers on grid-shaped sensor networks. The proposed approach could lead to the development of a sensor network that detects falls by performing CNN inference processing on the edge. We complement our work with a simulation using the simulator ns-3. The simulation is designed to emulate the communication overhead of the proposed approach applied to a wired sensor network that resembles the smart floor. Simulation results provide evidence on the feasibility of the proposed concept applied to wired grid shaped sensor networks.

关键词： Sensor networks edge computing fall detection convolutional neural networks network simulator ns-3

来源：评论

学校读者我要写书评

暂无评论

A Novel Integrated Simulation Framework for Cyber-Physical Systems Modelling

引用

IEEE TRANSACTIONS ON PARALLEL AND distributed SYSTEMS 2023年第10期34卷 2684-2698页

作者： Tampouratzis, Nikolaos Mousouliotis, Panagiotis Papaefstathiou, Ioannis Aristotle Univ Thessaloniki Lab Comp Syst Architecture Thessaloniki 54124 Greece

The growing use of Cyber-Physical Systems (CPS) in a plethora of domains (e.g. healthcare, industry, smart homes, transportation, etc.) triggers an urgent demand for simulation frameworks that can simulate in an integrated manner all the components (i.e. CPUs, Memories, networks, Physical Environment) of a system-under-design(SuD). By utilizing such a simulator, software design can proceed in parallel with physical development which results in the reduction of the so important time-to-market. The main problem, however, is that currently there is a shortage of such simulation frameworks;most simulators used for modelling the digital aspects of CPS applications (i.e. full-system CPU/Mem/Peripheral simulators) lack any support of the CPS physical aspects and vice versa. The presented fully-distributed simulation framework (APOLLON) is the first known open-source, high-performance simulator that can handle holistically complex CPSs including processors, peripherals, networks and physical aspects of them. APOLLON is an extension of the COSSIM simulation framework and it integrates, in a novel and efficient way, a combined processing and network simulator with the widely-used Ptolemy physical simulator, in a transparent way. Our highly integrated approach is further augmented with Machine Learning capabilities by implementing Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) recurrent neural networks in both the Cyber and Physical domains, enabling users to develop their complex recurrent neural networks significantly fast and accurately. APOLLON has been evaluated when executing a number of benchmarks and real-world use cases;the end results demonstrate that the presented approach has up to 99% accuracy in the reported SuD aspects.

关键词： Cyber-physical system simulator distributed systems simulator integrated simulator machine learning

来源：评论

学校读者我要写书评

暂无评论

Scalable training of trustworthy and energy-efficient predictive graph foundation models for atomistic materials modeling: a case study with HydraGNN

引用

JOURNAL OF SUPERCOMPUTING 2025年第4期81卷 1-48页

作者： Pasini, Massimiliano Lupo Choi, Jong Youl Mehta, Kshitij Zhang, Pei Rogers, David Bae, Jonghyun Ibrahim, Khaled Z. Aji, Ashwin M. Schulz, Karl W. Polo, Jorda Balaprakash, Prasanna Oak Ridge Natl Lab Computat Sci & Engn Div 1 Bethel Valley Rd Oak Ridge TN 37831 USA Oak Ridge Natl Lab Comp Sci & Math Div 1 Bethel Valley Rd Oak Ridge TN 37831 USA Oak Ridge Natl Lab Natl Ctr Computat Sci 1 Bethel Valley Rd Oak Ridge TN 37831 USA Lawrence Berkeley Natl Lab Comp Sci Dept 1 Cyclotron Rd Berkeley CA 94720 USA AMD Res Adv Micro Devices Austin TX USA Oak Ridge Natl Lab Comp & Computat Sci Directorate 1 Bethel Valley Rd Oak Ridge TN 37831 USA

We present our work on developing and training scalable, trustworthy, and energy-efficient predictive graph foundation models (GFMs) using HydraGNN, a multi-headed graph convolutional neural network architecture. HydraGNN expands the boundaries of graph neural network (GNN) computations in both training scale and data diversity. It abstracts over message passing algorithms, allowing both reproduction of and comparison across algorithmic innovations that define nearest-neighbor convolution in GNNs. This work discusses a series of optimizations that have allowed scaling up the GFMs training to tens of thousands of GPUs on datasets consisting of hundreds of millions of graphs. Our GFMs use multitask learning (MTL) to simultaneously learn graph-level and node-level properties of atomistic structures, such as energy and atomic forces. Using over 154 million atomistic structures for training, we illustrate the performance of our approach along with the lessons learned on two state-of-the-art US Department of Energy (US-DOE) supercomputers, namely the Perlmutter petascale system at the National Energy Research Scientific Computing Center and the Frontier exascale system at Oak Ridge Leadership Computing Facility. The HydraGNN architecture enables the GFM to achieve near-linear strong scaling performance using more than 2000 GPUs on Perlmutter and 16,000 GPUs on Frontier.

关键词： Machine learning Atomistic materials modeling distributed data parallelism Graph foundation models Graph neural networks Large-scale data processing for machine learning

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：