检索结果-内蒙古大学图书馆

22nd IEEE International Conference on Machine Learning and Applications, ICMLA 2023

作者： Das, Madhureeta Bagwe, Gaurav Pan, Miao Yuan, Xiaoyong Zhang, Lan Michigan Technological University Department of Electrical and Computer Engineering HoughtonMI United States University of Houston Department of Electrical and Computer Engineering HoustonTX United States College of Computing Michigan Technological University HoughtonMI United States

ISBN: (纸本)9798350345346

With the popularity of tiny devices based on microcontroller units, there is an urgent need to develop federated tiny learning to privately obtain a well-performed tiny model serving tiny devices. However, due to the limited capacity of tiny models, the fundamental difference between training deep neural networks and tiny neural networks makes existing federated learning designed for deep models ineffective in learning tiny models. Although prior tiny machine learning research successfully augments tiny models with enlarged architecture for improved capacity, such augmentation relies on a pre-known centralized dataset and thus cannot be used in federated settings. To fill this void, in this work, we propose an innovative federated tiny learning framework, FedTinyAug, to enable distributed tiny model augmentation. By taking advantage of the extra capability at larger participating devices, the server first constructs augmented models and distributes them to larger devices, providing auxiliary supervision for training the tiny model. To provide strong supervision, a gradient-based augmented model selection algorithm is designed to efficiently determine favorable augmented models to fully explore distinct or even heterogeneous on-device knowledge. Extensive experiments are conducted on three popular tiny models to validate the effectiveness of FedTinyAug. Key augmentation factors are evaluated to guide the implementation of FedTinyAug in practice. © 2023 IEEE.

关键词： Federated Learning Tiny Machine Learning

来源：评论

学校读者我要写书评

暂无评论

Constrained Reinforcement Learning Under Model Mismatch 41

Constrained Reinforcement Learning Under Model Mismatch

引用

41st International Conference on Machine Learning, ICML 2024

作者： Sun, Zhongchang He, Sihong Miao, Fei Zou, Shaofeng Department of Electrical Engineering University at Buffalo New York United States School of Computing University of Connecticut Storrs United States Department of Computer Science & Engineering University at Buffalo New York United States

Existing studies on constrained reinforcement learning (RL) may obtain a well-performing policy in the training environment. However, when deployed in a real environment, it may easily violate constraints that were originally satisfied during training because there might be model mismatch between the training and real environments. To address this challenge, we formulate the problem as constrained RL under model uncertainty, where the goal is to learn a policy that optimizes the reward and at the same time satisfies the constraint under model mismatch. We develop a Robust Constrained Policy Optimization (RCPO) algorithm, which is the first algorithm that applies to large/continuous state space and has theoretical guarantees on worst-case reward improvement and constraint violation at each iteration during the training. We show the effectiveness of our algorithm on a set of RL tasks with constraints. Copyright 2024 by the author(s)

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

A New Hybrid Transformer Topology for Distribution Networks 49

A New Hybrid Transformer Topology for Distribution Networks

引用

49th Annual Conference of the IEEE Industrial Electronics Society, IECON 2023

作者： Power, Rupert Madawala, Udaya Baguley, Craig Song, Bingkun The University of Auckland Department of Electrical Computer and Software Engineering Auckland New Zealand Auckland University of Technology Department of Engineering Computing and Mathematical Sciences Auckland New Zealand

ISBN: (纸本)9798350331820

The widespread integration of Distributed Generation (DG) and new loads such as Electric Vehicles (EVs) into power distribution networks presents substantial technical challenges for network operators, such as voltage fluctuations and phase unbalance. These can be addressed by many power electronic devices with advanced control strategies, including the Hybrid Transformer (HT). This paper presents a new three-phase HT concept which can perform multiple functions, including voltage regulation and voltage balancing, to address power quality problems in distribution networks. Theory and control strategy are presented, along with simulated results incorporating an unbalance-tolerant control scheme and 3D Space Vector Modulation, to demonstrate the validity of the proposed HT topology. Its advantages include being retrofittable to existing distribution transformers, not requiring bulky low-frequency converter magnetics, having a low semiconductor component count, and allowing reduced DC-link capacitances. This makes it cost-effective, compact and light weight in comparison to the existing topologies that provide similar functionality. © 2023 IEEE.

关键词： Topology

来源：评论

学校读者我要写书评

暂无评论

Sequential Patch Analysis Framework for Lung Disease Classification 11th

Sequential Patch Analysis Framework for Lung Disease Classi...

引用

11th International Conference on Future Data and Security engineering, FDSE 2024

作者： Le, Kim-Ngoc T. Le, Duc-Tai Choo, Hyunseung Department of AI Systems Engineering Sungkyunkwan University Seoul Korea Republic of Department of Electrical and Computer Engineering Sungkyunkwan University Seoul Korea Republic of College of Computing and Informatics Sungkyunkwan University Seoul Korea Republic of

ISBN: (纸本)9789819604364

Deep neural network (DNN) models have shown promise in lung disease classification using spectrogram transformations of respiratory sounds. However, these models typically process entire images, which can be computationally expensive, and they require large datasets for effective training. With the largest available respiratory dataset, ICBHI, containing only 920 recordings, training a robust DNN becomes challenging. To address this issue, this paper introduces a simplified DNN framework that incorporates novel techniques such as patch splitting and transfer learning. Specifically, the proposed framework splits the spectrogram image into a sequence of fixed-size patches, effectively reducing computational demands. These patches are processed with the position embeddings by DNN encoders to capture acoustic characteristics, leveraging a pre-trained ImageNet model to make efficient use of the limited dataset. Evaluation experiments on the ICBHI dataset show that the proposed method improves the state-of-the-art 8-class classification accuracy by 1.91%. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

关键词： Lung cancer

来源：评论

学校读者我要写书评

暂无评论

ETAG: An Energy-Neutral Ear Tag for Real-Time Body Temperature Monitoring of Dairy Cattle 23

ETAG: An Energy-Neutral Ear Tag for Real-Time Body Temperatu...

引用

29th Annual International Conference on Mobile computing and Networking, MobiCom 2023

作者： Vu, Hien Chung, Hanwook Choi, Christopher Kim, Younghyun Department of Electrical and Computer Engineering College of Engineering Wisconsin Embedded Systems and Computing Laboratory University of Wisconsin-Madison MadisonWI60032179 United States

ISBN: (纸本)9781450399906

Heat stress, caused by a warming climate and the increasingly high milk-producing dairy cattle, is one of the major threats to the well-being of dairy cattle as well as the economic, environmental, and social sustainability of dairy farming around the world. Timely identification of cows under heat stress is crucial to improving animal welfare, preventing milk production losses, and preserving water and energy for *** paper presents a smart ear tag, named eTag, and an associated system that can read a passive microchip temperature sensor subcutaneously injected into the animal with minimal discomfort. It features a lightweight design using a single coil shared for microchip scanning and wireless charging. eTag is autonomously recharged by a wireless charger over the head during daily milking sessions, enabling perpetual operation without battery replacement after deployment. The real-world performance of the proposed system was examined intensively in a three-week deployment on seven lactating Holstein cows. We demonstrate that eTag can reliably collect accurate body temperature in real time while maintaining a positive energy flow. The deployment of eTag will enable the timely detection of heat stress and facilitate precision control of barn cooling systems. © 2023 ACM.

关键词： Energy harvesting

来源：评论

学校读者我要写书评

暂无评论

Innovative fusion of VGG16, MobileNet, EfficientNet, AlexNet, and ResNet50 for MRI-based brain tumor identification

引用

Iran Journal of computer Science 2024年第1期8卷 185-215页

作者： Kia, Marjan Sadeghi, Soroush Safarpour, Homayoun Kamsari, Mohammadreza Jafarzadeh Ghoushchi, Saeid Ranjbarzadeh, Ramin Department of Business Information System Business Analysis (ISBA) Park University Parkville United States School of Electrical and Computer Engineering University of Tehran Tehran Iran Department of Software Engineering University of Szeged Szeged Hungary Faculty of Electrical Engineering Malek-Ashtar University of Technology(MUT) Esfahan Iran Faculty of Industrial Engineering Urmia University of Technology Urmia Iran School of Computing Faculty of Engineering and Computing Dublin City University Dublin Ireland

This study presents a novel approach for brain MRI classification by integrating multiple state-of-the-art deep learning (DL) architectures, including VGG16, EfficientNet, MobileNet, AlexNet, and ResNet50, with an attention mechanism. The primary objective was to accurately detect the presence or absence of brain tumors—a critical challenge in medical diagnosis. The methodology was validated on a comprehensive dataset comprising 3264 brain MRI scans, systematically divided into training, validation, and testing subsets. Preprocessing techniques tailored for MRI data were employed to enhance input quality. Each model was initially developed and evaluated individually to establish a performance baseline before their integration. The proposed ensemble technique capitalized on the complementary strengths of these models to enhance classification accuracy and robustness. The inclusion of an attention mechanism further refined the models by emphasizing critical regions in the MRI scans, thereby boosting predictive performance. In addition, data augmentation techniques were applied to mitigate overfitting and increase dataset resilience. Performance metrics, including precision, accuracy, and recall, demonstrated significant improvements across all evaluations, particularly for the models incorporating the attention mechanism. These findings highlight the potential of combining advanced DL architectures with attention mechanisms as a promising strategy for precise and reliable brain tumor classification in MRI scans. © The Author(s), under exclusive licence to Springer Nature Switzerland AG 2024.

关键词： AlexNet Attention mechanism Brain tumor classification MobileNet ResNet50

来源：评论

学校读者我要写书评

暂无评论

Federated Offline Reinforcement Learning: Collaborative Single-Policy Coverage Suffices 41

Federated Offline Reinforcement Learning: Collaborative Sing...

引用

41st International Conference on Machine Learning, ICML 2024

作者： Woo, Jiin Shi, Laixi Joshi, Gauri Chi, Yuejie Department of Electrical and Computer Engineering Carnegie Mellon University PittsburghPA15213 United States Department of Computing Mathematical Sciences California Institute of Technology CA91125 United States

Offline reinforcement learning (RL), which seeks to learn an optimal policy using offline data, has garnered significant interest due to its potential in critical applications where online data collection is infeasible or expensive. This work explores the benefit of federated learning for offline RL, aiming at collaboratively leveraging offline datasets at multiple agents. Focusing on finite-horizon episodic tabular Markov decision processes (MDPs), we design FedLCB-Q, a variant of the popular model-free Q-learning algorithm tailored for federated offline RL. FedLCB-Q updates local Q-functions at agents with novel learning rate schedules and aggregates them at a central server using importance averaging and a carefully designed pessimistic penalty term. Our sample complexity analysis reveals that, with appropriately chosen parameters and synchronization schedules, FedLCB-Q achieves linear speedup in terms of the number of agents without requiring high-quality datasets at individual agents, as long as the local datasets collectively cover the state-action space visited by the optimal policy, highlighting the power of collaboration in the federated setting. In fact, the sample complexity almost matches that of the single-agent counterpart, as if all the data are stored at a central location, up to polynomial factors of the horizon length. Furthermore, FedLCB-Q is communication-efficient, where the number of communication rounds is only linear with respect to the horizon length up to logarithmic factors. Copyright 2024 by the author(s)

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Dielectric Resonator Magnetoelectric Dipole Arrays with Low Cross Polarization,Backward Radiation,and Mutual Coupling for MIMO Base Station Applications

引用

China Communications 2024年第6期21卷 290-304页

作者： Song Simin Da Yiran Qian Bingyi Huang Xiaoyu Chen Xiaoming Li Yingsong Ahmed A.Kishk School of Information and Communications Engineering Xi’an Jiaotong UniversityXi’an 710049China Key Laboratory of Intelligent Computing and Signal Processing Ministry of Education Anhui UniversityHefei 230039China Department of Electrical and Computer Engineering Concordia UniversityMontrealQC H3G 1M8Canada

Dielectric resonator magnetoelectric dipole(DRMED)arrays with enhanced isolation,reduced cross-polarization,and backward radiation are proposed for base station(BS)*** proposed antenna comprises an elevated dielectric resonator antenna(DRA)on a small metal plate above a sizeable common ground *** DRA is designed in its T Eδ11 mode,acting like a magnetic *** surface current excited by the differential probes flowing on the small ground plane is equivalent to an electric *** these two equivalent dipoles are orthogonal,they have the magnetoelectric dipole characteristics with reduced backward ***,the small ground planes can be treated as decoupling structures to provide a neutralization path to cancel the original coupling path.A linearly-polarized 4-element prototype array was verified experimentally in previous ***,a dual-polarized DRMED antenna is presented to construct a 2-element and 4×4 array for BS *** investigate its MIMO performance,sophisticated multi-cell scenario simulations are carried *** using the proposed dualpolarized DRMED array,the cellular system capacityis improved by 118.6%compared to a conventional DRA *** significant MIMO system improvement is mainly due to the reduced backward radiation and,therefore,reduced inter-cell *** align well with the simulations.

关键词： backward radiation dielectric resonator antennas(DRAs) isolation magnetoelectric-dipole(ME-dipole)antennas

来源：评论

学校读者我要写书评

暂无评论

Lungcanary- Pioneering Early Lung Cancer Detection using Machine Learning Algorithms 8

Lungcanary- Pioneering Early Lung Cancer Detection using Mac...

引用

8th IEEE International Conference on Computational System and Information Technology for Sustainable Solutions, CSITSS 2024

作者： Govindarajan, Priya Abhishek, M. Niharika, K. Tesfahun, Abebe Amrita School of Computing Amrita Vishwa Vidyapeetham Department of Computer Science Mysuru Campus India School of Electrical & Computer Engineering Debre Markos Institute of Technology Debre Markos University Ethiopia Ethiopia

ISBN: (纸本)9798331505462

Lung cancer, a malignant tumour originating in the lung cells, is notoriously challenging to detect early through conventional clinical procedures, which often require invasive techniques. This study introduces LungCanary, an innovative approach leveraging machine learning algorithms to address the global health challenge of early lung cancer detection. By employing advanced computational methods and utilizing transfer learning principles, LungCanary effectively extracts meaningful features from medical imaging data. The model incorporates multiple decision-making components to capture diverse data patterns, significantly enhancing diagnostic precision. Through rigorous experimentation, LungCanary demonstrated superior performance with an accuracy of 98.36%, a precision of 96%, and an error rate of just 1.64%. These findings highlight LungCanary's potential to outperform existing models, marking a breakthrough in the accuracy and reliability of lung cancer diagnostics. The significance of this research lies in its potential to revolutionize early detection methodologies, ultimately improving patient outcomes. © 2024 IEEE.

关键词： Lung cancer

来源：评论

学校读者我要写书评

暂无评论

A Novel Cluster Analysis-Based Crop Dataset Recommendation Method in Precision Farming

引用

computer Systems Science & engineering 2023年第9期46卷 3239-3260页

作者： K.R.Naveen Kumar Husam Lahza B.R.Sreenivasa Tawfeeq Shawly Ahmed A.Alsheikhy H.Arunkumar C.R.Nirmala Department of Computer Science&Engineering Bapuji Institute of Engineering&TechnologyDavangereKarnatakaIndia Department of Information Technology Faculty of Computing and Information TechnologyKing Abdulaziz UniversityJeddahSaudi Arabia Department of Information Science&Engineering Bapuji Institute of Engineering&TechnologyDavangereKarnatakaIndia Department of Electrical Engineering Faculty of Engineering at RabighKing Abdulaziz UniversityJeddahSaudi Arabia Department of Electrical Engineering College of EngineeringNorthern Border UniversityArarSaudi Arabia

Data mining and analytics involve inspecting and modeling large pre-existing datasets to discover decision-making *** agriculture uses datamining to advance agricultural *** farmers aren’t getting the most out of their land because they don’t use precision *** harvest crops without a well-planned recommendation *** crop production is calculated by combining environmental conditions and management behavior,yielding numerical and categorical *** existing research still needs to address data preprocessing and crop categorization/***,statistical analysis receives less attention,despite producing more accurate and valid *** study was conducted on a dataset about Karnataka state,India,with crops of eight parameters taken into account,namely the minimum amount of fertilizers required,such as nitrogen,phosphorus,potassium,and pH *** research considers rainfall,season,soil type,and temperature parameters to provide precise cultivation recommendations for high *** presented algorithm converts discrete numerals to factors first,then reduces ***,the algorithm generates six datasets,two fromCase-1(dataset withmany numeric variables),two from Case-2(dataset with many categorical variables),and one from Case-3(dataset with reduced factor variables).Finally,the algorithm outputs a class membership allocation based on an extended version of the K-means partitioning method with lambda *** presented work produces mixed-type datasets with precisely categorized crops by organizing data based on environmental conditions,soil nutrients,and ***,the prepared dataset solves the classification problem,leading to a model evaluation that selects the best dataset for precise crop prediction.

关键词： Data mining crop prediction k-prototypes k-means cluster 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：