检索结果-内蒙古大学图书馆

41st International Conference on machine Learning, ICmL 2024

作者： Niu, Puhua Wu, Shili Fan, mingzhou Qian, Xiaoning Department of Electrical and Computer Engineering Texas A&M University College StationTX United States Department of Computer Science and Engineering Texas A&M University College StationTX United States Computational Science Intiative Brookhaven National Laboratory UptonNY United States

Generative Flow Networks (GFlowNets) have been shown effective to generate combinatorial objects with desired properties. We here propose a new GFlowNet training framework, with policy-dependent rewards, that bridges keeping flow balance of GFlowNets to optimizing the expected accumulated reward in traditional Reinforcement-Learning (RL). This enables the derivation of new policy-based GFlowNet training methods, in contrast to existing ones resembling value-based RL. It is known that the design of backward policies in GFlowNet training affects efficiency. We further develop a coupled training strategy that jointly solves GFlowNet forward policy training and backward policy design. Performance analysis is provided with a theoretical guarantee of our policy-based GFlowNet training. Experiments on both simulated and real-world datasets verify that our policy-based strategies provide advanced RL perspectives for robust gradient estimation to improve GFlowNet performance. Our code is available at: ***/niupuhua1234/GFN-PG. Copyright 2024 by the author(s)

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Single-Loop Stochastic Algorithms for Difference of max-Structured Weakly Convex Functions 38

Single-Loop Stochastic Algorithms for Difference of Max-Stru...

引用

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Hu, Quanqi Qi, Qi Lu, Zhaosong Yang, Tianbao Department of Computer Science & Engineering Texas A&M University United States Department of Computer Science The University of Iowa United States Department of Industrial and Systems Engineering University of Minnesota United States

In this paper, we study a class of non-smooth non-convex problems in the form of minx[maxy∈Y (x, y) − maxz∈Z ψ(x, z)], where both Φ(x) = maxy∈Y (x, y) and Ψ(x) = maxz∈Z ψ(x, z) are weakly convex functions, and (x, y), ψ(x, z) are strongly concave functions in terms of y and z, respectively. It covers two families of problems that have been studied but are missing single-loop stochastic algorithms, i.e., difference of weakly convex functions and weakly convex strongly-concave min-max problems. We propose a stochastic moreau envelope approximate gradient method dubbed SmAG, the first single-loop algorithm for solving these problems, and provide a state-of-the-art non-asymptotic convergence rate. The key idea of the design is to compute an approximate gradient of the moreau envelopes of Φ, Ψ using only one step of stochastic gradient update of the primal and dual variables. Empirically, we conduct experiments on positive-unlabeled (PU) learning and partial area under ROC curve (pAUC) optimization with an adversarial fairness regularizer to validate the effectiveness of our proposed algorithms. © 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

On optimal streaming kernelization algorithms

引用

science China(Information sciences) 2024年第8期67卷 337-338页

作者： Hao FENG Wei YANG Jianer CHEN School of Computer Science Guangzhou University Department of Computer Science and Engineering Texas A&M University

The streaming model has been a popular model in big data computation. Streaming kernelization algorithms can be regarded as data compression processes on streaming data. In this study, we give a general method for developing computational lower bounds for streaming kernelization algorithms that is applicable to a large class of computational problems. As an example, we use the method to prove computational lower bounds for the well-known problem d-Hi TTin GSET.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Automated Knowledge Framework for IoT Cybersecurity Compliance

Automated Knowledge Framework for IoT Cybersecurity Complian...

引用

2024 IEEE International Conference on Big Data, BigData 2024

作者： Oranekwu, Ikechukwu Elluri, Lavanya Batra, Gunjan Texas A&m University - Central Texas Dept. of Computer Information Systems Killeen United States Kennesaw State University Dept. of Information Systems and Security Kennesaw United States

ISBN: (纸本)9798350362480

Rapid expansion in the manufacture and use of Internet of Things (IoT) devices has introduced significant challenges in ensuring compliance with cybersecurity standards. To protect user data and privacy, all organizations providing IoT devices must adhere to complex guidelines such as the National Institute of Standards and Technology Inter agency Report (NIST IR) 8259, which defines essential cybersecurity guidelines for IoT manufacturers. However, interpreting and applying these rules from these guidelines and the privacy policies remains a significant challenge for companies. Thus, this project presents a novel approach to extract knowledge from NIST 8259 for creating semantically rich ontology mappings. Our ontology captures key compliance rules, which are stored in a knowledge graph (KG) that allows organizations to crosscheck and update privacy policy documents with ease. The KG also enables real-time querying using SPARQL and offers a transparent view of regulatory adherence for IoT manufacturers and users. By automating the process of verifying cybersecurity compliance, the framework ensures that companies remain aligned with NIST standards, eliminating manual checks and reducing the risk of non-compliance. We also demonstrate that compared to the baseline Large Language models (LLms), our proposed framework has more compliance accuracy, and is more efficient and scalable. © 2024 IEEE.

关键词： Ontology

来源：评论

学校读者我要写书评

暂无评论

Assessing the Effects of a Soft Passive Low-Back Exoskeleton for Emergency medical Services 4

Assessing the Effects of a Soft Passive Low-Back Exoskeleton...

引用

4th IEEE International Conference on Human-machine Systems, ICHmS 2024

作者： mukherjee, Tiash Rana Do Vale Gunter, Tiago Xavier manchanda, Eshan Tyagi, Oshin mehta, Ranjana K. Pagilla, Prabhakar Texas A&m University Dept. Of Mechanical Engineering College Station United States Texas A&m University School Of Public Health College Station United States Research University Of Michigan Dept. Of Industrial Engineering & Operations Ann-Arbor United States University Of Wisconsin Dept. Of Industrial Engineering Madison United States

ISBN: (纸本)9798350315790

This study examines the effectiveness of passive soft low back exoskeletons (LBEs) in mitigating low back injuries among emergency medical technicians and paramedics (EmT-Ps). To evaluate the dynamic interaction between users (EmT-Ps) and the ergonomic intervention (LBEs), perceptions of physical and mental exertion, muscle activity of the shoulder, lower back and lower limbs and range of motion (ROm) were investigated. The study involved a controlled field study with 20 EmT-Ps performing a modified physical agility test with and without the assistance of the LBE. The physical test circuital included tasks such as cardiac pulmonary resuscitation, stretcher pull-push, stair chair carry, etc., which are commonly performed by EmT-Ps in their occupation. EmT-Ps perceived lower physical demands associated with the circuital task when using the LBE, whereas no differences were observed with perceived mental demands. Physiological results show a reduction in muscular stress in the lower-back muscles for some tasks and comparable ROm for all tasks. Collectively, the study findings indicate that the LBEs are effective in some, but not all, emergency medical service tasks while not hindering naturalistic movements and work performance. © 2024 IEEE.

关键词： muscle

来源：评论

学校读者我要写书评

暂无评论

Decentralized AI-based Fault Detection and Localization to Enhance Dynamic Response of Grid-Forming Inverters 50

Decentralized AI-based Fault Detection and Localization to E...

引用

50th Annual Conference of the IEEE Industrial Electronics Society, IECON 2024

作者： Behnam, Reza Gohari, Amirhosein Shadmand, mohammad B. Bayhan, Sertac Abu-Rub, Haitham Dept. of Electrical & Computer Engineering University of Illinois Chicago IL United States Qatar Environment and Energy Research Institute Hamad Bin Khalifa University Doha Qatar Dept. of Electrical & Computer Engineering Texas A&M University at Qatar Doha Qatar

ISBN: (纸本)9781665464543

Grid-forming inverters (GFmIs) are promising solutions for voltage and frequency support in upcoming power electronics-dominated grids (PEDG). However, current state-of-the-art decentralized control schemes for GFmIs are designed for operation under normal conditions. These decentralized control schemes could result in an adverse dynamic response if a cluster of GFmIs network disconnects from the rest of the grid due to a fault. The dynamic response of GFmIs could be improved with coordinated control schemes which require communication and make the system more complex. This paper proposes a decentralized Artificial Intelligence (AI)-based method for online fault detection and localization to enhance the dynamic response of GFmIs without using a communication layer. Each GFmI will detect and localize the line tripping based on its output active and reactive power measurement. Then, according to the topology of the grid after line tripping, each GFmI updates its nominal power and inertia to suppress the frequency transient caused by the line tripping. The inertia and damping factor of the GFmIs are then recalculated to ensure optimal operation of the network of inverters after line tripping. Several case studies are presented to validate the effectiveness of the proposed method in the timely detection, localization, and mitigation of adverse frequency transient after line tripping in a communication-less manner. © 2024 IEEE.

关键词： Synchronous generators

来源：评论

学校读者我要写书评

暂无评论

Blockchain-Enhanced Framework for Secure Third-Party Vendor Risk management and Vigilant Security Controls

Blockchain-Enhanced Framework for Secure Third-Party Vendor ...

引用

2024 IEEE International Conference on Big Data, BigData 2024

作者： Gupta, Deepti Elluri, Lavanya Jain, Avi moni, Shafika Showkat Aslan, Omer Texas A&m University - Central Texas Dept. of Computer Information Systems TX United States College of Business University of Tampa FL United States Embry-Riddle Aeronautical University Dept. of Electrical Engineering and Computer Science FL United States Bandirma Onyedi Eylül University Department of Software Engineering Bandirma Turkey

ISBN: (纸本)9798350362480

In an era of heightened digital interconnectedness, businesses increasingly rely on third-party vendors to enhance their operational capabilities. However, this growing dependency introduces significant security risks, making it crucial to develop a robust framework to mitigate potential vulnerabilities. This paper proposes a comprehensive secure framework for managing third-party vendor risk, integrating blockchain technology to ensure transparency, traceability, and immutability in vendor assessments and interactions. By leveraging blockchain, the framework enhances the integrity of vendor security audits, ensuring that vendor assessments remain up-to-date and tamperproof. This proposed framework leverages smart contracts to reduce human error while ensuring real-time monitoring of compliance and security controls. By evaluating critical security controls - such as data encryption, access control mechanisms, multi-factor authentication, and zero-trust architecture - this approach strengthens an organization's defense against emerging cyber threats. Additionally, continuous monitoring enabled by blockchain ensures the immutability and transparency of vendor compliance processes. In this paper, a case study on iHealth's transition to AWS Cloud demonstrates the practical implementation of the framework, showing a significant reduction in vulnerabilities and marked improvement in incident response times. Through the adoption of this blockchain-enabled approach, organizations can mitigate vendor risks, streamline compliance, and enhance their overall security posture. Our findings highlight the importance of employing blockchain to enforce security controls and maintain compliance with healthcare regulations such as HIPAA. In this paper, we present a comprehensive set of security controls and demonstrate how blockchain technology enhances their effectiveness, ensuring greater transparency, accountability, and automation in vendor assessments. By reducing human error, e

关键词： Smart contract

来源：评论

学校读者我要写书评

暂无评论

AI-Driven Proportionate Power Sharing in Virtual Synchronous Generators for Optimizing the Source Conditions and Efficiency 50

AI-Driven Proportionate Power Sharing in Virtual Synchronous...

引用

50th Annual Conference of the IEEE Industrial Electronics Society, IECON 2024

作者： Asif, Uzair D’silva, Silvanus Shadmand, mohammad B. Bayhan, Sertac Abu-Rub, Haitham Dept. of Electrical & Computer Engineering University of Illinois Chicago IL United States Qatar Environment and Energy Research Institute Hamad Bin Khalifa University Doha Qatar Dept. of Electrical & Computer Engineering Texas A&M University at Qatar Doha Qatar

ISBN: (纸本)9781665464543

Virtual impedance-based power sharing between several interconnected grid-forming inverters (GFmI)s in a power electronics-dominated grid (PEDG) does not consider the available power reserves. This non-optimal power-sharing may result in the overloading of nearby inverter-based resources (IBR)s exhibiting lower effective line impedances in the network. This paper proposes an optimal and controllable power-sharing scheme based on a modular artificial neural network (ANN) that can consider various factors like power reserve, line losses, incentives from the grid, etc., and increases the system's scalability. An optimum power allocator (OPA) mechanism that monitors the output active powers and line losses of all the IBRs is proposed to address the issue of uncontrolled power sharing. The OPA dynamically adjusts power allocation weights using linear programming to optimize the overall loss minimization cost function. moreover, an artificial neural network-based virtual impedance optimizer (ANN-VIO) is also proposed, which estimates the corresponding virtual impedance (VI) value for each GFmI unit in the PEDG. The primary controller of GFmI uses these estimated VI values to adjust its point of common coupling (PCC) voltage reference and attain optimum proportionate power sharing within the overall PEDG. multiple case studies are performed to validate the effectiveness of the proposed GFmIs power-sharing scheme. © 2024 IEEE.

关键词： Cost functions

来源：评论

学校读者我要写书评

暂无评论

Accelerated Primal-Dual Projection Neurodynamic Approach With Time Scaling for Linear and Set Constrained Convex Optimization Problems

引用

IEEE/CAA Journal of Automatica Sinica 2024年第6期11卷 1485-1498页

作者： You Zhao Xing He mingliang Zhou Tingwen Huang Chongqing Key Laboratory of Nonlinear Circuits and Intelligent Information Processing College of Electronic and Information EngineeringSouthwest UniversityChongqing 400715China School of Computer Science Chongqing UniversityChongqing 400044China Department of Mathematics Texas A&M University at QatarDoha 23874Qatar IEEE

The Nesterov accelerated dynamical approach serves as an essential tool for addressing convex optimization problems with accelerated convergence *** previous studies in this field have primarily concentrated on unconstrained smooth con-vex optimization *** this paper,on the basis of primal-dual dynamical approach,Nesterov accelerated dynamical approach,projection operator and directional gradient,we present two accelerated primal-dual projection neurodynamic approaches with time scaling to address convex optimization problems with smooth and nonsmooth objective functions subject to linear and set constraints,which consist of a second-order ODE(ordinary differential equation)or differential conclusion system for the primal variables and a first-order ODE for the dual *** satisfying specific conditions for time scaling,we demonstrate that the proposed approaches have a faster conver-gence *** only requires assuming convexity of the objective *** validate the effectiveness of our proposed two accel-erated primal-dual projection neurodynamic approaches through numerical experiments.

关键词： Accelerated projection neurodynamic approach lin-ear and set constraints projection operators smooth and nonsmooth convex optimization time scaling.

来源：评论

学校读者我要写书评

暂无评论

Early Diagnosis of mild Cognitive Impairment Using Deep Neural Network with Single and multi-input Analysis 11

Early Diagnosis of Mild Cognitive Impairment Using Deep Neur...

引用

11th IEEE International Conference on Healthcare Informatics, ICHI 2023

作者： Abbasian, Pouneh Zargarnezhad, Hanif Hammond, Tracy Multidisciplinary Engineering Texas A&m University College StationTX United States Texas A&m University Mechanical Engineering College StationTX United States Texas A&m University Computer Science and Engineering College StationTX United States

ISBN: (纸本)9798350302639

Convolutional Neural Networks (CNNs) have made significant advancements in the diagnosis of hard-to-detect illnesses, such as mild Cognitive Impairment (mCI) and its prodromal stage Alzheimer's Disease (AD). However, identifying early mCI (EmCI) remains a challenge since the brain patterns of EmCI and Normal Cognitive (NC) individuals are quite similar. Recent studies have shown that CNNs are effective in analyzing multiple neuroimaging datasets. However, more work remains to be done on multi-input (imaging and non-imaging) CNN models. By leveraging models based on a shallow CNN and hybrid neural network, this paper investigates if the accuracy of diagnosing EmCI could be enhanced using multi-input data (e.g., structural magnetic Resonance Imaging (smRI), demographic data). These models have fewer parameters while retaining competitive performance on EmCI vs. NC classification problems, requiring less computational effort. Experiments were performed on a publicly available Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset with 546 subjects, including 164 EmCI and 282 NC from ADNI3 and 50 EmCI and 50 NC from ADNI2. The hybrid model achieved 87% accuracy in distinguishing between NC and EmCI groups using a coronal view of the brain. © 2023 IEEE.

关键词： magnetic resonance imaging

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：