检索结果-内蒙古大学图书馆

15th International Conference on Information and Communication technology Convergence, ICTC 2024

作者： Lee, Hyunsoo Kim, Joongheon Korea University Department of Electrical and Computer Engineering Seoul Korea Republic of

ISBN: (纸本)9798350364637

This paper explores the application of reinforcement learning (RL) to torpedo guidance with a focus on obstacle avoidance and target acquisition in dynamic environments. By employing a dual-actor network approach and a regularization technique for the critic network, the proposed method demonstrates robust performance in scenarios with random obstacle placements and target directions. Our results show that the system effectively navigates toward the target while avoiding obstacles, leveraging reinforcement learning to optimize torpedo trajectories. The study addresses the challenges of overestimation and underestimation biases common in reinforcement learning methods, such as deep deterministic policy gradient (DDPG), by utilizing the double actors and regularized critics (DARC) algorithm. This approach has potential implications for enhancing the autonomy and efficiency of military applications in complex and uncertain settings. © 2024 IEEE.

关键词： Torpedoes

来源：评论

学校读者我要写书评

暂无评论

Real-Time Health Monitoring System by Integrating Augmented Reality and Internet of Things 6th

Real-Time Health Monitoring System by Integrating Augmented ...

引用

6th International Conference on Intelligent Communication, Control and Devices, ICICCD 2024

作者： Kulkarni, Radhika V. Shilimkar, Avanish Bhegade, Sakshi Dadmal, Jidnyasa Department of Computer Engineering Vishwakarma Institute of Technology Pune India

ISBN: (纸本)9789819783281

In order to provide real-time health monitoring in the healthcare industry, the article presents a novel integration of augmented reality (AR) and the Internet of Things (IoT). Our goal is to build a unified system that combines IoT and AR technologies to produce an interesting health monitoring service. Our study reveals the combined potential of these technologies, which have been explored separately in previous research. We close the gap in the literature by utilising AR for vital indicator monitoring and demonstration, which allows for accurate real-time patient monitoring. Our system architecture incorporates Blynk Cloud for data storage, Unity 3D for complex augmented reality health data visualisation, MQTT protocol for secure data sharing, and sensors connected to ESP32s. Our method, which improves medical education and technology integration, lays the groundwork for a revolutionary change in healthcare by utilising ARKit, ARCore, Vuforia, and OpenCV. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Electronic health record

来源：评论

学校读者我要写书评

暂无评论

SmartAgroChain: Revolutionizing Agricultural Supply Chains with Blockchain technology 1

SmartAgroChain: Revolutionizing Agricultural Supply Chains w...

引用

1st IEEE International Conference on Building a Secure and Empowered Cyberspace, BuildSEC 2024

作者： Sarker, Shuvro Israq, Mobashera Abdal, Labib Mim, Umma Hani Sourov, Nahian Islam, Salekul Department of Computer Science and Engineering United International University Dhaka Bangladesh Department of Electrical and Computer Engineering North South University Dhaka Bangladesh

ISBN: (纸本)9798331516727

Agricultural crop and its supply chains are a fundamental component of our life and means of subsistence, influencing the lives of farmers and consumers alike and guaranteeing the continuous flow of important food supplies. It is plagued with inefficiencies, a lack of transparency and unfair pricing, all of which have an adverse effect on both farmers and consumers. This article introduces SmartAgroChain, a blockchain-based decentralized strategy, which will transform the agricultural supply chain by ensuring equality, trust and transparency among all parties involved. In this system, farmers publish their produced products together with their predicted prices, which starts a community-driven bidding or auction process with retailers. All price related activities take place through blockchain transactions. Our research results demonstrate that SmartAgroChain successfully removes intermediaries, provides fair pricing and builds users' confidence. In comparison to traditional marketplaces, field implementations in semi-urban and rural areas exhibit significant adoption rates and transaction volumes. Both farmers and consumers gain from the competitive price that is ensured by the open bidding procedure. Moreover, we have developed a escrow system for payment that removes the fraudulent activities. We have designed the necessary algorithms for bidding and payment processes and implemented through Ethereum smart contracts. Our performance study shows feasibility the scalability of SmartAgroChain system. © 2024 IEEE.

关键词： Smart contract

来源：评论

学校读者我要写书评

暂无评论

ACHIEVING SAMPLE AND COMPUTATIONAL EFFICIENT REINFORCEMENT LEARNING BY ACTION SPACE REDUCTION VIA GROUPING 12

ACHIEVING SAMPLE AND COMPUTATIONAL EFFICIENT REINFORCEMENT L...

引用

12th International Conference on Learning Representations, ICLR 2024

作者： Li, Yining Ju, Peizhong Shroff, Ness Department of Electrical and Computer Engineering United States Department of Computer Science and Engineering The Ohio State University ColumbusOH43210 United States

Reinforcement learning often needs to deal with the exponential growth of states and actions when exploring optimal control in high-dimensional spaces (often known as the curse of dimensionality). In this work, we address this issue by learning the inherent structure of action-wise similar MDP to appropriately balance the performance degradation versus sample/computational complexity. In particular, we partition the action spaces into multiple groups based on the similarity in transition distribution and reward function, and build a linear decomposition model to capture the difference between the intra-group transition kernel and the intra-group rewards. Both our theoretical analysis and experiments reveal a surprising and counter-intuitive result: while a more refined grouping strategy can reduce the approximation error caused by treating actions in the same group as identical, it also leads to increased estimation error when the size of samples or the computation resources is limited. This finding highlights the grouping strategy as a new degree of freedom that can be optimized to minimize the overall performance loss. To address this issue, we formulate a general optimization problem for determining the optimal grouping strategy, which strikes a balance between performance loss and sample/computational complexity. We further propose a computationally efficient method for selecting a nearly-optimal grouping strategy, which maintains its computational complexity independent of the size of the action space. © 2024 12th International Conference on Learning Representations, ICLR 2024. 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

HETEROGENEOUS PERSONALIZED FEDERATED LEARNING BY LOCAL-GLOBAL UPDATES MIXING VIA CONVERGENCE RATE 12

HETEROGENEOUS PERSONALIZED FEDERATED LEARNING BY LOCAL-GLOBA...

引用

12th International Conference on Learning Representations, ICLR 2024

作者： Jiang, Meirui Le, Anjie Li, Xiaoxiao Dou, Qi Department of Computer Science and Engineering The Chinese University of Hong Kong Hong Kong Department of Electrical and Computer Engineering The University of British Columbia Canada

Personalized federated learning (PFL) has emerged as a promising technique for addressing the challenge of data heterogeneity. While recent studies have made notable progress in mitigating heterogeneity associated with label distributions, the issue of effectively handling feature heterogeneity remains an open question. In this paper, we propose a personalization approach by Local-Global updates Mixing (LG-Mix) via Neural Tangent Kernel (NTK)-based convergence. The core idea is to leverage the convergence rate induced by NTK to quantify the importance of local and global updates, and subsequently mix these updates based on their importance. Specifically, we find the trace of the NTK matrix can manifest the convergence rate, and propose an efficient and effective approximation to calculate the trace of a feature matrix instead of the NTK matrix. Such approximation significantly reduces the cost of computing NTK, and the feature matrix explicitly considers the heterogeneous features among samples. We have theoretically analyzed the convergence of our method in the over-parameterize regime, and experimentally evaluated our method on five datasets. These datasets present heterogeneous data features in natural and medical images. With comprehensive comparison to existing state-of-the-art approaches, our LG-Mix has consistently outperformed them across all datasets (largest accuracy improvement of 5.01%), demonstrating the outstanding efficacy of our method for model personalization. Code is available at https://***/med-air/HeteroPFL. © 2024 12th International Conference on Learning Representations, ICLR 2024. All rights reserved.

关键词： Mixing

来源：评论

学校读者我要写书评

暂无评论

Fault-tolerant visual servo control for a robotic arm with actuator faults

引用

Neural Computing and Applications 2024年第25期36卷 15815-15828页

作者： Li, Jiashuai Peng, Xiuyan Li, Bing Sreeram, Victor Wu, Jiawei College of Intelligent Systems Science and Engineering Harbin Engineering University Heilongjiang Nantong Harbin150001 China Department of Electrical Electronic and Computer Engineering The University of Western Australia Stirling Highway PerthWA6009 Australia

The study targets uncertain coupling faults in robotic arm actuators and proposes a new fault-tolerant visual servo control strategy. Specifically, it considers both multiplicative and additive actuator faults within the dynamic of the robotic arm, treating the coupling faults and time-varying disturbances as an aggregate of concentrated uncertainties. A radial basis function neural network-based state observer is introduced to online approximate these concentrated uncertainties, which include fault information, eliminating the need for prior knowledge of faults. Furthermore, a fault-tolerant controller based on a non-singular fast terminal sliding mode is proposed, which separately decouples the nominal quantities and concentrated uncertainties and develops individual adaptive control laws for each. This effectively reduces the detrimental impact of coupled faults and disturbances on the system’s performance, facilitating image feature trajectory tracking control with minimal jitter, high precision, and strong transient response capabilities. The stability of the state observer and the fault-tolerant controller has been substantiated through Lyapunov’s theory. Lastly, numerical simulations validate the efficacy and robustness of the proposed fault-tolerant visual servo control approach. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2024.

关键词： Robotic arms

来源：评论

学校读者我要写书评

暂无评论

Cloud-Edge Framework for Multi-Vehicle Deployment Control using Spiking Neural Networks

Cloud-Edge Framework for Multi-Vehicle Deployment Control us...

引用

2024 IEEE Research and Applications of Photonics in Defense Conference, RAPID 2024

作者： Banad, Yaser Mike Sharif, Sarah Safura The University of Oklahoma Department of Electrical and Computer Engineering Oklahoma United States

ISBN: (纸本)9798350373097

Recent advances have spotlighted the use of unmanned aerial vehicles (UAVs) for defense and surveillance, focusing on cooperative monitoring and exploration. However, these UAV systems face significant operational challenges due to limited computational and energy resources, affecting their endurance and effectiveness in critical missions. This study proposes a novel approach using Spiking Neural Networks (SNNs) to overcome these limitations. SNNs, inspired by the human brain's processing efficiency, offer a shift towards more autonomous and adaptive UAV systems. Leveraging SNNs, we introduce a cloud-edge computational framework aimed at optimizing barrier coverage control and enhancing UAVs' monitoring capabilities within resource constraints. This SNN-based framework marks a considerable advancement in UAV operational capabilities, enabling longer missions, better area coverage, and adaptability to environmental changes. The effectiveness in achieving optimal barrier coverage sets a new benchmark for multivehicle systems in defense applications, underscoring the potential of brain-inspired computing in military operations and paving the way for future research in autonomous defense technologies. © 2024 IEEE.

关键词： Neural networks

来源：评论

学校读者我要写书评

暂无评论

A 1.19GHz 9.52Gsamples/sec Radix-8 FFT Hardware Accelerator in 28nm

A 1.19GHz 9.52Gsamples/sec Radix-8 FFT Hardware Accelerator ...

引用

2024 IEEE Hot Chips 36 Symposium, HCS 2024

作者： Tang, Larry Chen, Siyuan Harisrikanth, Keshav Xu, Guanglin Franchetti, Franz Mai, Ken Carnegie Mellon University Department of Electrical and Computer Engineering United States

Dedicated FFT hardware is typically designed as a standalone block and then integrated into systems Relatively fixed functionality once implemented System-level integration with FFT-based application software becomes ...

ISBN: (纸本)9798350388503

关键词：

来源：评论

学校读者我要写书评

暂无评论

Intelligent Handover Management Enabled by O-RAN and Deep Reinforcement Learning 100

Intelligent Handover Management Enabled by O-RAN and Deep Re...

引用

100th IEEE Vehicular technology Conference, VTC 2024-Fall

作者： Dai, Jiongyu Mahboob, Shadab Wang, Haining Liu, Lingjia Virginia Tech Bradley Department of Electrical and Computer Engineering United States

ISBN: (纸本)9798331517786

The increased number of connected devices and diverse quality of service demands in cellular networks present formidable challenges for efficient traffic management. In this paper, we present a handover control approach to tackle these challenges in 5G networks, integrating the Proximal Policy Optimization (PPO) framework with Reservoir Computing (RC). Our approach, aligned with the O-RAN architecture, offers a near-real-time solution for intelligent handover management, improving cell performance, and enhancing user experience. This renders our solution O-RAN compliant and deployable as an 'xApp' on the RAN Intelligent Controller (RIC). The introduced algorithm features a sequential state design including user-specific and cell-specific metrics. Harnessing the power of reservoir computing, our solution captures dynamic changes in state sequences while reducing training overhead and enhancing online training efficiency. The efficacy of our algorithm is demonstrated through evaluations on the ns3 platform, proving its superiority in system-level simulations. The experimental results affirm the effectiveness of the introduced algorithm in addressing the dynamic challenges of handover in cellular networks. © 2024 IEEE.

关键词： Deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

An Efficient Text Cleaning Pipeline for Clinical Text for Transformer Encoder Models 12

An Efficient Text Cleaning Pipeline for Clinical Text for Tr...

引用

12th IEEE International Conference on Intelligent Systems, IS 2024

作者： Ridoy, Shahriyar Zaman Sultana, Jannat Ria, Zinnat Fowzia Uddin, Mohammed Arif Rahman, Md Hasibur Rahman, Rashedur M. North South University Department of Electrical and Computer Engineering Dhaka Bangladesh

ISBN: (纸本)9798350350982

It might be challenging to choose the best text preprocessing strategy in the field of natural language processing (NLP) due to the variety of techniques available. Given the popularity of transformer models, we wondered if preprocessing was necessary and, if so, what methods would improve the models' performance. Especially when working with clinical text data, accuracy is crucial. Our goal was to find an appropriate pre-processing pipeline for clinical texts that maintains or improves model performance. We experienced four common preprocessing techniques and their groupings on two datasets from MIMIC-3 and PubMed. We used four models: BERT base, BioBERT, BioClinicalBERT, and RoBERTa. The varied accuracy results from existing techniques inspired us to develop a new pipeline to improve accuracy. Our pipeline starts with removing repeated punctuation, normalizing the text with a CleanText function, and filtering less important words using TF-IDF scores to keep clinically applicable terms and moderate noise. Our results presented that our pipeline outperformed the base models. For the MIMIC-3 dataset, the BERT base model achieved 90.16% accuracy, and for the PubMed dataset, BioBERT achieved 64.20% accuracy. We also found that removing stop words decreased accuracy, while using TF-IDF either maintained or improved it up to 3%. Additionally, as we removed less important words from the documents our pipeline considerably reduced training time up to 17%. © 2024 IEEE.

关键词： Natural language processing systems

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：