检索结果-内蒙古大学图书馆

Towards Fair and Scalable Trial Assignment in Federated Bandits: A Shapley Value Approach

IEEE Transactions on Big Data 2024年 1-14页

作者： Wang, Zibo Zhu, Yifei Wang, Dan Han, Zhu UM-SJTU Joint Institute Shanghai Jiao Tong University Shanghai China UM-SJTU Joint Institute and Cooperative Medianet Innovation Center Shanghai Jiao Tong University Shanghai China Department of Computing The Hong Kong Polytechnic University Hong Kong China Department of Electrical and Computer Engineering University of Houston Houston USA

Federated multi-armed bandits extend the multi-armed bandits framework to the federated learning setting where multiple clients in the same exploration space collaboratively identify the optimal arm. While previous studies demonstrated its efficiency like classical federated learning, in this paper, we present the first work that reveals the serious fairness problem in federated multi-armed bandits when clients have heterogeneous and overlapping armsets. The fairness problem happens because clients with different trial requirements should conduct different numbers of trials summing up to a global requirement, but they wish to minimize their exploration effort. To address the novel fairness concern, we formally formulate the fairness-aware trial assignment as a coalitional game. Based on the theoretically derived trial requirements, we devise a Shapley value-based trial assignment mechanism to guarantee fairness. Regardless of the #P-hard complexity when deriving the general Shapley value, we achieve an accurate computation of trial assignment with polynomial complexity by exploiting its unique characteristic. We further carefully control the numbers of trials in each iteration to resolve the communication bottleneck and minimize the wasted trials. Experiment results show that, compared to the naïve federated scheme, our design outperforms with both high fairness metrics and high efficiency in total trials and communication. IEEE

关键词： Investments

来源：评论

学校读者我要写书评

暂无评论

Automating life science labs at the single-cell level through precise ultrasonic liquid sample ejection:PULSE

引用

Microsystems & Nanoengineering 2024年第6期10卷 173-188页

作者： Peiran Zhang Zhenhua Tian Ke Jin Kaichun Yang Wesley Collyer Joseph Rufo Neil Upreti Xianjun Dong Luke P.Lee Tony Jun Huang Department of Mechanical Engineering and Materials Science Duke UniversityDurhamNC 27708USA Department of Mechanical Engineering Virginia Polytechnic Institute and State UniversityBlacksburgVA 24061USA Genomics and Bioinformatics Hub Department of NeurologyBrigham and Women’s HospitalHarvard Medical SchoolBostonMAUSA Department of Medicine Brigham and Women’s HospitalHarvard Medical SchoolBostonMAUSA Department of Bioengineering Department of Electrical Engineering and Computer ScienceUniversity of California at BerkeleyBerkeleyCAUSA Institute of Quantum Biophysics Department of BiophysicsSungkyunkwan UniversitySuwonKorea Department of Chemistry&Nanoscience Ewha Womans UniversitySeoulKorea

Laboratory automation technologies have revolutionized biomedical ***,the availability of automation solutions at the single-cell level remains scarce,primarily owing to the inherent challenges of handling cells with such small dimensions in a precise,biocompatible ***,we present a single-cell-level laboratory automation solution that configures various experiments onto standardized,microscale test-tube matrices via our precise ultrasonic liquid sample ejection technology,known as *** enables the transformation of titer plates into microdroplet arrays by printing nanodrops and single cells acoustically in a programmable,scalable,and biocompatible *** pipetting robots,PULSE enables researchers to conduct biological experiments using single cells as anchoring points(e.g.,1 cell vs.1000 cells per“tube”),achieving higher resolution and potentially more relevant data for modeling and downstream *** demonstrate the ability of PULSE to perform biofabrication,precision gating,and deterministic array barcoding via preallocated droplet-addressable *** cells can be gently printed at a speed range of 5–20 cell⋅s−1 with an accuracy of 90.5–97.7%,which can then adhere to the substrate and grow for up to 72 h while preserving cell *** the deterministic barcoding experiment,95.6%barcoding accuracy and 2.7%barcode hopping were observed by comparing the phenotypic data with known genotypic data from two types of single *** PULSE platform allows for precise and dynamic analyses by automating experiments at the single-cell level,offering researchers a powerful tool in biomedical research.

关键词： precise liquid offering

来源：评论

学校读者我要写书评

暂无评论

Net-Zero Smart Home Monitoring System

Net-Zero Smart Home Monitoring System

引用

Conference on Technologies for Sustainability (Sustech)

作者： Sarkis A. Sherian Tamer R. Omar Sunai Kim Electrical and Computer Engineering Department California State Polytechnic University Pomona Civil Engineering Deaprtment California State Polytechnic University Pomona

ISBN: (数字)9798331504311

ISBN: (纸本)9798331504328

Many innovations in the renewable energy sector are being explored for deployment at residential and commercial sites with cost effectiveness and sustainability in mind. Although various solutions exist that allow for local monitoring and control of individual peripherals, there still exists a need for a truly integrated whole home solution for existing residential housings to convert them into smart homes without major upfront investment costs all the while still delivering the complete solution of control and monitoring of energy consuming peripherals. This paper and the corresponding project will validate the feasibility of such system by utilizing various cloud computing resources and building four cyber-physical subsystems that interactively communicate with the cloud to offer a solution suitable for any home. The prototypes will focus on implementing control and monitoring of various sensors and peripherals in a typical home that can improve optimization of energy consumption and improve efficiency by making autonomous decisions as well as educating consumers of energy utilization habits. A single node power consumption monitoring device will be implemented to mimic a smart branch circuit breaker where all non-critical loads can be connected. A second subsystem will detect a room's occupancy status. The third subsystem is a window position sensor that uploads a window open or close status to the cloud. The fourth subsystem will emulate a closed loop hot water recirculating pump system with active feedback on cold point temperature and real time indication of full circulation of hot water within water pipes. An M5Stack Core2 ESP32 IoT development kit for Amazon Web Services (AWS) will be utilized as the primary controller interface that will interact with Amazon Web Services cloud computing resources and enable communication and control of various internet-of-things (IoT) cyber-physical subsystems. Various AWS resources will be utilized to acquire, monitor, and

关键词： Temperature sensors Temperature measurement Cloud computing Renewable energy sources Circuit breakers Net zero Smart homes Windows Internet of Things Monitoring

来源：评论

学校读者我要写书评

暂无评论

SVM Modeling Simulation to Evaluate the Electric Vehicle Transmitting Points 2

SVM Modeling Simulation to Evaluate the Electric Vehicle Tra...

引用

2nd IEEE International Conference on Advances in Computing, Communication and Applied Informatics, ACCAI 2023

作者： Kaushal, Rajanish Kumar Agal, Sanjay Neeraja, B. Singh, Ravinjit Singh, Pavitar Parkash Chandigarh University Department of Electrical Engineering Punjab Mohali India Dr. V. R. Godhania College of Engineering and Technology Department of Computer Engineering Gujarat Porbandar India Government Polytechnic College Department of Electrical and Electronics Engineering Hyderabad India Malaysia Lovely Professional University Department of Management India

ISBN: (纸本)9798350315905

Green energy-based intelligent grids are needed to improve security, operation conditions, and power management. Different sources, like solar, wind turbines etc., generate green *** green energy will reduce pollution and improves energy production. The current research uses the machine learning model to apply green energy management in an intelligent grid by smart monitoring. The existing Support vector model will predict the need for hybrid electric vehicle (HEV) charging requirements. Coordinate and innovative/intelligent charging systems are applicable in HEVs. The dragonfly-based model is used to evaluate the best charging system for optimization purposes. Apart from this, the self-adaptive model is used to get modified or suit the best charging strategy. Simulation results obtained from the intelligent microgrid reveal the model's suitability and efficiency. By the end of the research, predict the charging requirements concerning minor errors and compare the coordinate and smart charging system performance and operational cost. © 2023 IEEE.

关键词： Support vector machines

来源：评论

学校读者我要写书评

暂无评论

Development of a Machine Learning Model For Big Data Analytics

Development of a Machine Learning Model For Big Data Analyti...

引用

2023 International Conference on Science, engineering and Business for Sustainable Development Goals, SEB-SDG 2023

作者： Musa, Ugbedeojo Adebiyi, Marion O. Adebiyi, Abayomi A. Adebiyi, Avodele A. Federal Polytechnic Idah Department of Computer Science Kogi Nigeria Landmark University Department of Computer Science Kwara Umu-Aran Nigeria Faculty of Engineering and the Built Environment Department of Electrical Power Engineering Durban South Africa University of Technology Durban South Africa

ISBN: (纸本)9798350324785

The aim of the study was to utilize machine learning approaches in the development of a big data analytics model that can effectively discover patterns and aid in making intelligent decisions. The impact of Big Data on scientific discoveries and technology in Big Data analytics cannot be overstated. In this regard, the study focused on the development of a system to control incoming illegal messages from intruders/unknown users, in order to alleviate illicit contextual communication channels. Unstructured data sets were collected from online blogs and labeled to suit the needs of the machine learning algorithm. The data was preprocessed using Weka machine learning libraries and converted into attribute related file form (arff). The resampling technique was used to partition the data into 80% training set and 20% testing set, which was used to develop the big data model and identify illicit message content to prevent people from being victimized. The Na ï ve Bayes machine learning algorithm was utilized to predict the categorized labels as binary values. The results of the study showed that Naïve Bayes had an accuracy of 96.76% for metrics evaluation of true precision model. It is recommended that the application of other machine learning techniques to the classification model would improve its performance. © 2023 IEEE.

关键词： Big data

来源：评论

学校读者我要写书评

暂无评论

A Wireless Synchronized System for the Long-Term Global Health Monitoring 7

A Wireless Synchronized System for the Long-Term Global Heal...

引用

7th International Conference on Smart and Sustainable Technologies, SpliTech 2022

作者： Coviello, Giuseppe Avitabile, Gianfranco Talarico, Claudio Wang-Roveda, Janet M. Florio, Antonello Polytechnic University of Bari Department of Electrical and Information Engineering BA Bari70125 Italy Gonzaga University Department of Electrical and Computer Engineering SpokaneWA99258 United States University of Arizona Department of Electrical and Computer Engineering TucsonAZ85721 United States

ISBN: (纸本)9789532901160

The advent of the Internet of Medical Things allows using sensors implanted or placed over the body for monitoring people's global health. However, classical problems like network time synchronization are still present and need to be solved. In this paper, we present a wireless synchronized platform for long-term off-line monitoring of global health through the acquisition of heart rate, peripheral oxygen saturation, and gait and posture analysis based on muscle activity assessment. The network time synchronization is achieved thanks to a low-power accurate time synchronization algorithm developed by the authors. Experimental results on a healthy subject prove both the accuracy of time synchronization and the proposed system functionality. © 2022 University of Split, FESB.

关键词： Synchronization

来源：评论

学校读者我要写书评

暂无评论

A Behavioral Model for Transcutaneous Carbon Dioxide Measurement Using Time-Correlated Single Photon Counting Technique

A Behavioral Model for Transcutaneous Carbon Dioxide Measure...

引用

Annual International Conference of the IEEE engineering in Medicine and Biology Society (EMBC)

作者： Isil Isiksalan Hakan B. Karli Tuna B. Tufan John Matthews Bige D. Unluturk Ulkuhan Guler Department of Electrical and Computer Engineering Worcester Polytechnic Institute Worcester MA USA Department of Electrical and Computer Engineering Michigan State University Michigan MI USA

ISBN: (数字)9798350371499

ISBN: (纸本)9798350371505

Millimeter-size carbon dioxide (CO 2 )-sensitive luminescence films emerge as practical sensors in miniaturized biomedical devices for measuring transcutaneous CO 2 . The intensity and lifetime of the excited luminophores in the film carry critical information about the CO 2 values in the surrounding environment. It has been proven that lifetime-based measurement is superior to intensity-based measurement because of its robustness to confounding factors. Addressing the challenge of accurately determining the rapid nanosecond-scale lifetime of CO 2 -sensitive luminophores, we introduce a behavioral model using a time-correlated single photon counting system equipped with silicon photomultipliers. We included nonidealities such as dark count rate, afterpulsing, and crosstalk to this model. To recover the lifetime from photon counts, we applied denoising and fitting techniques. With hyperparameter optimization to determine various post-processing parameters, we improved accuracy in recovery for lifetime values ranging from 3 ns to 7 ns. Our results demonstrate the applicability of the proposed model and techniques in real-world scenarios.

关键词： Semiconductor device modeling Accuracy Films Carbon dioxide Biomedical measurement Silicon Sensors System-on-chip Photonics Single-photon avalanche diodes

来源：评论

学校读者我要写书评

暂无评论

BOURNE: Bootstrapped Self-Supervised Learning Framework for Unified Graph Anomaly Detection 40

BOURNE: Bootstrapped Self-Supervised Learning Framework for ...

引用

40th IEEE International Conference on Data engineering, ICDE 2024

作者： Liu, Jie He, Mengting Shang, Xuequn Shi, Jieming Cui, Bin Yin, Hongzhi School of Computer Science Northwestern Polytechnical University Xi'an China The Hong Kong Polytechnic University Department of Computing Hong Kong Hong Kong School of Computer Science Peking University Beijing China School of Electrical Engineering & Computer Science The University of Queensland Brisbane Australia

ISBN: (纸本)9798350317152

Graph anomaly detection (GAD) has gained increasing attention in recent years due to its critical application in a wide range of domains, such as social networks, financial risk management, and traffic analysis. Existing GAD methods can be categorized into node and edge anomaly detection models based on the type of graph objects being detected. However, these methods typically treat node and edge anomalies as separate tasks, overlooking their associations and frequent co-occurrences in real-world graphs. As a result, they fail to leverage the complementary information provided by node and edge anomalies for mutual detection. Additionally, state-of-the-art GAD methods, such as CoLA and SL-GAD, heavily rely on negative pair sampling in contrastive learning, which incurs high computational costs, hindering their scalability to large graphs. To address these limitations, we propose a novel unified graph anomaly detection framework based on bootstrapped self-supervised learning (named BOURNE). We extract a subgraph (graph view) centered on each target node as node context and transform it into a dual hypergraph (hypergraph view) as edge context. These views are encoded using graph and hypergraph neural networks to capture the representations of nodes, edges, and their associated contexts. By swapping the context embeddings between nodes and edges and measuring the agreement in the embedding space, we enable the mutual detection of node and edge anomalies. Furthermore, BOURNE can eliminate the need for negative sampling, thereby enhancing its efficiency in handling large graphs. Extensive experiments conducted on six benchmark datasets demonstrate the superior effectiveness and efficiency of BOURNE in detecting both node and edge anomalies. © 2024 IEEE.

关键词： Anomaly detection

来源：评论

学校读者我要写书评

暂无评论

On Neural Collapse in Contrastive Learning with Imbalanced Datasets

On Neural Collapse in Contrastive Learning with Imbalanced D...

引用

IEEE Workshop on Machine Learning for Signal Processing

作者： Thuan Nguyen Ruijie Jiang Shuchin Aeron Prakash Ishwar D. Richard Brown Department of Engineering Engineering Technology and Surveying East Tennessee State University Department of Electrical and Computer Engineering Tufts University Medford Department of Electrical and Computer Engineering Boston University Boston MA Department of Electrical and Computer Engineering Worcester Polytechnic Institute Worcester MA

ISBN: (数字)9798350372250

ISBN: (纸本)9798350372267

Neural collapse is a phenomenon in neural networks where all the samples from the same class collapse to their class mean and the class means have a specific geometric structure called Equiangular Tight Frame (ETF). Recent studies have empirically and theoretically confirmed that neural collapse solutions attain a global minimum of Contrastive Learning (CL) losses if the class distribution is uniform (balanced classes). This paper investigates the neural collapse phenomenon in Contrastive Learning for imbalanced datasets. We show that even if the classes are imbalanced, global minima of Contrastive Learning losses include solutions in which all samples from the same class collapse to their class mean. However, the geometric structure of the class means is, in general, different from an ETF. In addition, we show that under certain conditions the optimal geometry of the class means can be found by solving a convex optimization problem. We provide a program based on the CVX package to find these optimal class means.

关键词： Geometry Conferences Neural networks Contrastive learning Machine learning Signal processing Convex functions

来源：评论

学校读者我要写书评

暂无评论

Three-way trade-off in multi-objective learning: optimization, generalization and conflict-avoidance

The Journal of Machine Learning Research

引用

The Journal of Machine Learning Research 2024年第1期25卷 9153-9205页

作者： Lisha Chen Heshan Fernando Yiming Ying Tianyi Chen Department of Electrical Computer & Systems Engineering Rensselaer Polytechnic Institute School of Mathematics and Statistics University of Sydney NSW Australia

Multi-objective learning (MOL) often arises in machine learning problems when there are multiple data modalities or tasks. One critical challenge in MOL is the potential conflict among different objectives during the optimization process. Recent works have developed various dynamic weighting algorithms for MOL, where the central idea is to find an update direction that avoids conflicts among objectives. Albeit its appealing intuition, empirical studies show that dynamic weighting methods may not outperform static ones. To understand this theory-practice gap, we focus on a stochastic variant of MGDA, the Multi-objective gradient with Double sampling (MoDo), and study the generalization performance and its interplay with optimization through the lens of algorithmic stability in the framework of statistical learning theory. We find that the key rationale behind MGDA--updating along conflict-avoidant direction--may hinder dynamic weighting algorithms from achieving the optimal O(1/√n) population risk, where n is the number of training samples. We further demonstrate the impact of dynamic weights on the three-way trade-off among optimization, generalization, and conflict avoidance unique in MOL. We showcase the generality of our theoretical framework by analyzing other algorithms under the framework. Experiments on various multi-task learning benchmarks are performed to demonstrate the practical applicability. Code is available at https://***/heshandevaka/Trade-Off-MOL.

关键词： multi-objective optimization statistical learning theory algorithm stability Pareto stationarity gradient conflict

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：