检索结果-内蒙古大学图书馆

Privet: A Privacy-Preserving Vertical Federated Learning Service for Gradient Boosted Decision Tables

IEEE Transactions on Services Computing 2023年第5期16卷 3604-3620页

作者： Zheng, Yifeng Xu, Shuangqing Wang, Songlei Gao, Yansong Hua, Zhongyun Harbin Institute of Technology School of Computer Science and Technology Guangdong Shenzhen518055 China Csiro Data61 SydneyNSW2720 Australia

Vertical federated learning (VFL) has recently emerged as an appealing distributed paradigm empowering multi-party collaboration for training high-quality models over vertically partitioned datasets. Gradient boosting has been popularly adopted in VFL, which builds an ensemble of weak learners (typically decision trees) to achieve promising prediction performance. Recently there have been growing interests in using decision table as an intriguing alternative weak learner in gradient boosting, due to its simpler structure, good interpretability, and promising performance. In the literature, there have been works on privacy-preserving VFL for gradient boosted decision trees, but no prior work has been devoted to the emerging case of decision tables. Training and inference on decision tables are different from that in the case of generic decision trees, not to mention gradient boosting with decision tables in VFL. In light of this, we design, implement, and evaluate Privet, the first system framework enabling privacy-preserving VFL service for gradient boosted decision tables. Privet delicately builds on lightweight cryptography and allows an arbitrary number of participants holding vertically partitioned datasets to securely train gradient boosted decision tables. Extensive experiments over several real-world datasets and synthetic datasets demonstrate that Privet achieves promising performance, with utility comparable to plaintext centralized learning. © 2008-2012 IEEE.

关键词： Decision trees

来源：评论

学校读者我要写书评

暂无评论

First-principles investigation of adhesion strength and interfacial bonding in Mg/X(X=Ti,Zr,Hf,V,Nd,Cr,Mo,Mn,and Fe)interface

引用

Journal of Magnesium and Alloys 2025年第1期13卷 83-89页

作者： Longke Bao Zhifu Yao Yuhui Zhang Cuiping Wang Kaihong Zheng Rongpei Shi Xingjun Liu Fusheng Pan State Key Laboratory of Advanced Welding and Joining Harbin Institute of TechnologyShenzhen 518055China College of Materials Science and Engineering and Institute of Materials Genome and Big DataHarbin Institute of TechnologyShenzhen 518055China Institute of New Materials Guangdong Academy of SciencesGuangzhou 510650China College of Materials Science and Engineering Xiamen UniversityXiamen 361005China College of Materials Science and Engineering Chongqing UniversityChongqing 400045China

Magnesium alloys have gained extensive applications across various industries,including aerospace,transportation,and civil construction,owing to their excellent combinations of high specific strength and stiffness[1].However,their lim-ited strength due to the lack of effective strengthening phases has hindered their broader industrial applications[2].Never-theless,it has been challenging to achieve significant strength-ening due to the restricted solubility of alloying elements in magnesium[3].Thus,more and more efforts have been made to explore the concept of secondary phase-reinforced magne-sium alloys[2,4,5],where the secondary phase acts as re-inforcing agents within the magnesium matrix,resembling a composite material.

关键词： adhesion strength interfacial bonding aluminum alloys industrial applications solubility strengthening phases alloying elements Mg X interfaces

来源：评论

学校读者我要写书评

暂无评论

Reinforcement Learning in Cognitive Robots for Autonomous Path Planning 5

Reinforcement Learning in Cognitive Robots for Autonomous Pa...

引用

5th International Conference on data Intelligence and Cognitive Informatics, ICDICI 2024

作者： Das, Archana Poornima, R. Indu Deepa, G. Gokila Sheela, K. Selva Akshya, J. Choudhry, Mani Deepak New Horizon College of Engineering Department of Information Science and Engineering Bengaluru India KGiSL Institute of Technology Department of Information Technology Coimbatore India Ppg Institute of Technology Department of Artificial Intelligence and Data Science Coimbatore India KGiSL Institute of Technology Department of Artificial Intelligence and Data Science Coimbatore India Srm Institute of Science and Technology Department of Computational Intelligence Chennai India Srm Institute of Science and Technology Department of Computing Technologies Chennai India

ISBN: (纸本)9798350389609

Cognitive robots are intelligent systems that learn from their environment, adapt to dynamic changes, and make decisions without a human's direct intervention. This research studies the integration of Reinforcement Learning in cognitive robots with Q-Learning for autonomous path planning. The study addresses the most critical issues in traditional A∗ and Dynamic Programming algorithms by devising path-planning techniques, mainly the lack of adaptability and computational efficiency in real-time and unpredictable environments. Our approach is based on Q-Learning, which is a model-free RL algorithm allowing the robots to find paths autonomously by optimizing their paths and avoiding obstacles. The Q-Learning algorithm provides the possibility of learning optimal policies for the robot through iterative interaction with its environment balanced between exploration and exploitation of the decision process. A reward system is utilized by the proposed model, encouraging the robot to explore shorter, collision-free paths and adjust based on feedback in real time from its surroundings. The model was found to be significantly better than its counterparts. In this context, as indicated, the Q-Learning model runs at 11.5 seconds, faster than A∗ at 18.3 seconds and Dynamic Programming at 16.7 seconds, yet with an accuracy of 94% and was found to have the highest collision avoidance rate at 98%. Additionally, the adaptability of the model about the environment presents a marked difference in terms of path length optimization, having done so with a mean path length of 12.3, compared to approaches or models. The robustness and scalability of the Q-Learning model make it highly applicable in real-world applications. © 2024 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Diagnosing Dementia Disorder Detection Using an Improved Eliminate Particle Swarm Optimization (IEPSO) Based on Convolutional Neural Networks

引用

SN Computer science 2024年第6期5卷 695页

作者： Duraipandian, Kavitha Ambigapathy, Murugan Department of Computer Science and Engineering SRM Institute of Science and Technology Ramapuram Campus Tamil Nadu Chennai India Department of Data Science and Business Systems SRM Institute of Science and Technology SRM Nagar Tamil Nadu Kattankulathur Chennai India

Dementia is an un-repairable and continuous disease that affects a person's mental health. Symptoms of Dementia may vary from one person to another, there are no effective treatments for detecting Dementia. Machine learning algorithm has proven to be effectively used by many researchers in assisting optimal detection of Dementia. So, this research introduces an improved eliminate particle swarm optimization (IEPSO) based Convolutional Neural Network (IEPSO-CNN) for diagnosing Dementia without causing severe effects effectively. IEPSO is used to find suitable parameters while training the model to improve Dementia detection accuracy. This IEPSO is used for tuning the parameter of the two-level Visual Geometry Group 16 (VGG 16) model. The integration of IEPSO and two hierarchical layers of VGG 16 increases the model's Dementia detection performance. The IEPSO-CNN model performance is evaluated using 6400 Dementia and non-Dementia images. The performance analysis shows that the IEPSO-CNN obtained a maximum accuracy of 99% for detecting the dementia stages. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2024.

关键词： Convolutional neural network Deep learning Dementia detection Improved eliminate particle swarm optimization VGG 16

来源：评论

学校读者我要写书评

暂无评论

Real-Time Object Detection for an Intelligent Retail Checkout System 3

Real-Time Object Detection for an Intelligent Retail Checkou...

引用

3rd International Conference on Intelligent and Innovative Technologies in Computing, Electrical and Electronics, IITCEE 2025

作者： Vaneeta, M. Mugada, Swathi Patil, Prachi Bhausaheb Preethi, V.J. Deeksha, K. M.S.Ramaiah Institute of Technology Department of Artificial Intelligence & Data Science Bangalore India

ISBN: (纸本)9798331515911

Customers are made to wait for extended hours in busy shopping settings as traditional limits of barcode and RFID-based billing systems have a predisposition towards errors and inefficiencies in scanning untagged or weighted items. The methods that are in use today can't keep up with the demand and lead to queues at check-out counters causing customer discontentment. This paper presents an idea of a novel Smart Billing Counter, based on Object Detection using advanced deep learning, YOLOv8, and OCR technologies for smooth billing processes. It captures the image of items placed at the counter, identifies what it is with minimal error, extracts necessary information from labels concerning weighted items, and generates a bill automatically. The unique capability of this solution is to recognize identical items in one frame with total cost by quantity to optimize checkout times and minimize the possibility of errors or fraud. At an inference speed of 37.2ms per image with 93.6% accuracy, this significantly enhances operational efficiency and ensures that grocery retailers provide their customers with smooth, error-free, and swift checkout processes. This innovation is likely, therefore, to bring about a transformation in retail billing systems. It will enhance customer satisfaction and operations efficiency quite highly. © 2025 IEEE.

关键词： Sales

来源：评论

学校读者我要写书评

暂无评论

Identifying and Classifying Scientific Unknowns in Clinical Literature 18

Identifying and Classifying Scientific Unknowns in Clinical ...

引用

18th International Society for Knowledge Organization Conference, ISKO 2024

作者： Wenjing, Zhao Jintao, Ma Jian, Du National Institute of Health Data Science Peking University China Institute of Medical Technology Health Science Center Peking University China Department of Biomedical Engineering College of Future Technology Peking University China National Institute of Health Data Science Peking University China

ISBN: (纸本)9783987400469

The scientific unknowns embedded within academic literature serve as a crucial resource for illuminating the trajectory of scientific progress and facilitating the formulation of future research directions. Clinical medicine, the discipline intricately intertwined with the art of diagnosis and treatment, necessitates a profound comprehension of the uncertainties inherent in scientific knowledge at the disease level. Using PubMed as the data source and leveraging the optimal machine learning model, a comprehensive framework and methodology for the identification and classification of scientific unknowns is developed. Moreover, by utilizing Medical Subject Headings (MeSH) to map clinical literature and disease burden into a unified disease classification system, the study extends its purview to the disease level. Consequently, it unravels a preference towards addressing "novel scientific problems" over the pursuit of intricate challenges. An empirical analysis of the relationship between scientific unknowns and disease burden is further conducted, indicating diseases with higher levels of scientific unknowns often bearing greater burdens. © 2024 International Society for Knowledge Organization. All rights reserved.

关键词： Diagnosis

来源：评论

学校读者我要写书评

暂无评论

Arc plasma-deposited Co single-atom catalysts supported on an aligned carbon nanofiber for hydrogen peroxide electrosynthesis and an electro-Fenton process

引用

Carbon Energy 2024年第11期6卷 234-255页

作者： Chang‐Kyu Hwang Sooyeon Kim Ki Ro Yoon Thao Thi Le Chinh V.Hoang Jae Won Choi Wenjun Zhang Sae Yane Paek Chung Hyeon Lee Ji Hyun Lee Keun Hwa Chae Sohee Jeong Seung Yong Lee Byeong-Kwon Ju Sang Hoon Kim Sang Soo Han Jong Min Kim Materials Architecturing Research Center Korea Institute of Science and Technology (KIST))seculSecrngbuk-guRepublic of Korea Department of Micro/Nano Systems Korea UniversitySeoulSeongbuk-guRepublic of Korea Computational Science Research Center Korea Institute of Science and Technology (KIST)SeculSecrngbuk-guRepublic of Korea Advanced Textile R&D Department Korea Institute of Industrial Technology (KITECH)AmsanGyeonggi-doRepublic of Korea Extreme Materials Research Center Korea Institute of Science and Technology (KIST)SeoulSeongbuk:guRepubic of Korea Nanoscience and Technology KIST School University of Science and TechnologySeoulRepublic of Korea Advanced Analysis and Data Center Korea Institute of Science and TechnologySeoulSeongbuk-guRepublic of Korea Display and Nanosensor Laboratory School of Electrical Engineering Korea UniversitySeoulSeongbuk guRepublic of Korea KHU-KIST Department of Converging Science and Technology Kyung Hee UniversitySeoulSeongbuk-guRepubic of Korea

Atomically dispersed single-atom catalysts(SACs)on carbon supports show great promise for H_(2)O_(2) electrosynthesis,but conventional wet chemistry methods using particulate carbon blacks in powder form have limited their potential as two-electron(2^(e−))oxygen reduction reaction(ORR)***,we demonstrate high-performance Co SACs supported on a free-standing aligned carbon nanofiber(CNF)using electrospinning and arc plasma deposition(APD).Based on the surface oxidation treatment of aligned CNF and precise control of the deposition amount in a dry-based APD process,we successfully form densely populated Co SACs on aligned *** experimental analyses and density functional theory calculations,we reveal that Co SAC has a Co–N_(2)–O_(2) moiety with one epoxy group,leading to excellent 2^(e−)ORR ***,the aligned CNF significantly improves mass transfer in flow cells compared to randomly oriented CNF,showing an overpotential reduction of 30 mV and a 1.3-fold improvement(84.5%)in Faradaic efficiency,and finally achieves an outstanding production rate of 15.75 mol gcat^(−1) h^(−1) at 300 mA cm^(−2).The high-performance Co SAC supported on well-aligned CNF is also applied in an electro-Fenton process,demonstrating rapid removal of methylene blue and bisphenol F due to its exceptional 2e^(−)ORR activity.

关键词： arc plasma deposition electro-Fenton flow reactor hydrogen peroxide production oxygen reduction reaction single-atom catalysts

来源：评论

学校读者我要写书评

暂无评论

Blockchain driven Evidence Management System 3

Blockchain driven Evidence Management System

引用

3rd International Conference on Artificial Intelligence and Signal Processing, AISP 2023

作者： Mehta, Shyam Kumari, K. Shantha Jain, Paras Raikwar, Harshal Gore, Shubham SRM Institute of Science & Technology Data Science and Business Systems Chennai India SRM Institute of Science & Technology Computer Science and Engineering Chennai India

ISBN: (纸本)9798350320749

When a cognizable violation like murder, abduction, rape, theft, etc. is committed, a victim or someone acting on their behalf must submit an electronic first information report (e-FIR) to the police station. Due to the centralized nature of the e-FIR database, it is possible for the offense's record to be hacked, and it is also possible for fake e-FIRs to be purposefully registered. data transparency and integrity are therefore major issues with the e-FIR database. Indian government launched nation-wide Crime and Criminal Tracking Network and Systems (CCTNS) during 2009 and it is an efficient e-governance system. This paper provides a blockchain solution to handle the complaints given on both cognizable and non-cognizable complaints. Using real-world examples from previous events, the technology and security underlying the use of blockchain will be discussed. Police will file an e-FIR, which will be validated by the authorities, and on acceptance of the FIR, it will be encrypted and stored as a hash along with the timestamp and hash of the next block. Now, how blockchain makes it secure is that it doesn't let anybody make changes to the FIR without proof of work and a vote of consensus in which a majority of the blockchain must agree to the change. The hash will be stored in smart contracts using Ethereum. Our findings demonstrate a trade-off between the number of transactions contained in a single block on the blockchain ledger and the security level of various hashing algorithms for the offence data. © 2023 IEEE.

关键词： Blockchain

来源：评论

学校读者我要写书评

暂无评论

Advancing Alzheimer's Diagnosis Through Transfer Learning with Deep MRI Analysis

Advancing Alzheimer's Diagnosis Through Transfer Learning wi...

引用

2024 International Conference on Advances in Modern Age Technologies for Health and Engineering science, AMATHE 2024

作者： Adityasai, B. Babu, Nirmal Varghese Karunya Institute Of Technology And Sciences Division Of Computer Science And Engineering Coimbatore India Karunya Institute Of Technology And Sciences Division Of Data Science And Cyber Security Coimbatore India

ISBN: (纸本)9798350371567

Alzheimer's disease (AD) poses a significant health-care challenge, necessitating early and accurate diagnosis for effective management. This study presents a novel approach aimed at enhancing AD diagnosis through the integration of transfer learning, deep learning techniques, and meticulous data preprocessing methodologies. Leveraging a rich dataset from the Open Access Series of Imaging Studies (OASIS), comprising MRI scans from both healthy individuals and those with AD, the proposed method seeks to identify subtle abnormalities indicative of the disease. The methodology encompasses several key steps. Initially, MRI data undergoes rigorous cleaning and standardization using histogram equalization to ensure compatibility and remove irrelevant information. Subsequently, a two-step feature extraction process involving an autoencoder neural network is employed to identify significant AD-related features. These features are then combined with other relevant data to provide a comprehensive understanding of the disease. To optimize model efficiency, three algorithms iteratively refine the feature set, focusing on the most informative features associated with AD. Finally, a pretrained deep learning model, VGG19, is fine-tuned using the AD-specific dataset. Results demonstrate a significant improvement in diagnostic accuracy post-hyperparameter tuning, with accuracy increasing from 76.5% to 89.5%. The enhanced model performance across various metrics underscores the potential of the proposed approach to revolutionize AD diagnosis, offering promise for earlier detection and more accurate prognostication. This research contributes to advancing the field of neuroimaging-based diagnostic techniques for neurodegenerative diseases like Alzheimer's. © 2024 IEEE.

关键词： Magnetic resonance imaging

来源：评论

学校读者我要写书评

暂无评论

Deep Transfer Learning Based Multi-Classification of Dental Diseases X-Ray Images 4

Deep Transfer Learning Based Multi-Classification of Dental ...

引用

4th International Conference on Ubiquitous Computing and Intelligent Information Systems, ICUIS 2024

作者： Narmadha, R.P. Karthik, B. Mithiloshini, B. Saranya, R. KIT-KalaignarKarunanidhi Institute of Technology Dept. Artificial intelligence and data science Coimbatore India

ISBN: (数字)9798331529635

ISBN: (纸本)9798331529635

This research work proposes a deep transfer learning model for multi-classification purposes within dental diseases classed by their X-ray images. The paper elaborates typical challenges posed by dentistry, which faces many diagnosis issues;the decisive process of proper classification into types of diseases requires being fine and efficient. Most state-of-the-art methodologies developed through CNNs and DNNs have been found to yield fair results, but they both suffer from overfitting and are time-consuming processes. The new model uses the ResNet50 to get over those limitations. Dental X-ray pictorials are to be grouped into four distinct categories using the proposed architecture: cavities, fillings, impacted teeth, and implants. The system proposed should classify these four categories in X-ray images with an accuracy greater than 92%. Therefore, the proposed system will automatically identify areas requiring clinical treatment. Expected outcome is a strong and dependable tool for enhanced dental disease diagnosis by better accuracy, efficiency, and performance in the evvaluation of X-ray pictures. © 2024 IEEE.

关键词： Convolutional neural networks

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：