检索结果-内蒙古大学图书馆

Proceedings of the 36th International Conference on Neural information Processing systems

作者： Markus Hiller Mehrtash Harandi Tom Drummond School of Computing and Information Systems The University of Melbourne Department of Electrical and Computer Systems Engineering Monash University

ISBN: (纸本)9781713871088

Inspired by the concept of preconditioning, we propose a novel method to increase adaptation speed for gradient-based meta-learning methods without incurring extra parameters. We demonstrate that recasting the optimisation problem to a non-linear least-squares formulation provides a principled way to actively enforce a well- conditioned parameter space for meta-learning models based on the concepts of the condition number and local curvature. Our comprehensive evaluations show that the proposed method significantly outperforms its unconstrained counterpart especially during initial adaptation steps, while achieving comparable or better overall results on several few-shot classification tasks - creating the possibility of dynamically choosing the number of adaptation steps at inference time.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Reinforcement Learning Driven Smart Charging Algorithms to Enhance Battery Lifespan and Grid Sustainability

Reinforcement Learning Driven Smart Charging Algorithms to E...

引用

2024 International Conference on Advances in Modern Age Technologies for Health and engineering Science, AMATHE 2024

作者： Sivakumar, S. Yadav, Mali Ram Bharathi, A. Akila, D. Dineshkumar, P. Banupriya, V. Vel Tech Rangarajan Dr.Sagunthala R And D Institute Of Science And Technology Department Of Electrical And Electronics Engineering Chennai India Lovely Professional University Department Of Physical Education Punjab India Vels Institute Of Science Technology And Advanced Studies Department Of Information Technology Pallavaram India Saveetha College Of Liberal Arts And Sciences Simats Department Of Computer Applications Chennai India Sona College Of Technology Department Of Information Technology Tamil Nadu Salem India M. Kumarasamy College Of Engineering Department Of Computer Science And Business Systems Karur India

ISBN: (纸本)9798350371567

This study investigates the use of RL-driven smart charging algorithms, particularly the Deep Deterministic Policy Gradients (DDPG) model, to maximise grid sustainability and battery lifetime. The current state of this field's study is lacking in real-time adaptation to different settings and in handling the ever-changing nature of charging profiles. By comparison, the suggested approach demonstrates a significant efficiency boost through the use of RL principles, which allow the model to learn and adjust its charging tactics. In specifically, the strategy contributes to a 10% improvement in total grid efficiency while cutting peak demand by 25%, minimising charge-discharge cycles by 20%, and reducing battery degradation rates by 15%. Not only do these measurements show that the suggested method works, but they also show how unique it is in offering a smart charging solution that is both sensitive and adaptive. This research fills in the gaps left by other studies and proposes an optimisation technique that is more dynamic and works in real-time. Its influence on the health of individual batteries and the sustainability of the grid as a whole is noteworthy. © 2024 IEEE.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Blockchain-Empowered Lifecycle Management for AI-Generated Content (AIGC) Products in Edge Networks

arXiv

引用

arXiv 2023年

作者： Liu, Yinqiu Du, Hongyang Niyato, Dusit Kang, Jiawen Xiong, Zehui Miao, Chunyan Shen, Xuemin Jamalipour, Abbas The School of Computer Science and Engineering Nanyang Technological University Singapore The School of Automation Guangdong University of Technology China The Pillar of Information Systems Technology and Design Singapore University of Technology and Design Singapore The Department of Electrical and Computer Engineering University of Waterloo Canada The School of Electrical and Information Engineering University of Sydney Australia

The rapid development of Artificial Intelligence-Generated Content (AIGC) has brought daunting challenges regarding service latency, security, and trustworthiness. Recently, researchers presented the edge AIGC paradigm, effectively optimize the service latency by distributing AIGC services to edge devices. However, AIGC products are still unprotected and vulnerable to tampering and plagiarization. Moreover, as a kind of online non-fungible digital property, the free circulation of AIGC products is hindered by the lack of trustworthiness in open networks. In this article, for the first time, we present a blockchain-empowered framework to manage the lifecycle of edge AIGC products. Specifically, leveraging fraud proof, we first propose a protocol to protect the ownership and copyright of AIGC, called Proof-of-AIGC. Then, we design an incentive mechanism to guarantee the legitimate and timely executions of the funds-AIGC ownership exchanges among anonymous users. Furthermore, we build a multi-weight subjective logic-based reputation scheme, with which AIGC producers can determine which edge service provider is trustworthy and reliable to handle their services. Through numerical results, the superiority of the proposed approach is demonstrated. Last but not least, we discuss important open directions for further research. © 2023, CC BY.

关键词： Blockchain

来源：评论

学校读者我要写书评

暂无评论

Rethinking generalization in few-shot classification 22

Rethinking generalization in few-shot classification

引用

Proceedings of the 36th International Conference on Neural information Processing systems

作者： Markus Hiller Rongkai Ma Mehrtash Harandi Tom Drummond School of Computing and Information Systems The University of Melbourne Department of Electrical and Computer Systems Engineering Monash University

ISBN: (纸本)9781713871088

Single image-level annotations only correctly describe an often small subset of an image's content, particularly when complex real-world scenes are depicted. While this might be acceptable in many classification scenarios, it poses a significant challenge for applications where the set of classes differs significantly between training and test time. In this paper, we take a closer look at the implications in the context of few-shot learning. Splitting the input samples into patches and encoding these via the help of Vision Transformers allows us to establish semantic correspondences between local regions across images and independent of their respective class. The most informative patch embeddings for the task at hand are then determined as a function of the support set via online optimization at inference time, additionally providing visual interpretability of 'what matters most' in the image. We build on recent advances in unsupervised training of networks via masked image modelling to overcome the lack of fine-grained labels and learn the more general statistical structure of the data while avoiding negative image-level annotation influence, aka supervision collapse. Experimental results show the competitiveness of our approach, achieving new state-of-the-art results on four popular few-shot classification benchmarks for 5-shot and 1-shot scenarios.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Empowering secure and sustainable healthcare through federated learning and blockchain synergies in a Medical Internet of Things

引用

International Journal of Machine Learning and Cybernetics 2025年 1-33页

作者： Wang, Jing Manna, Sarbajit Aksoy, Muammer Sarkar, Arindam Rahman, Md Arafatur Noorwali, Abdulfattah Othman, Kamal M. Alenazi, Mohammed J. F. School of Computer and Information Qiannan Normal University for Nationalities Duyun China Institute for Big Data Analytics and Artificial Intelligence (IBDAAI) Universiti Teknologi MARA Shah Alam Malaysia Department of Computer Science and Electronics Ramakrishna Mission Vidyamandira Howrah India Cyber Security Department College of Sciences Al-Mustaqbal University Babylon Iraq Computer Information Systems Department Ahmed Bin Mohammed Military College Doha Qatar School of Mathematics and Computer Science University of Wolverhampton Wolverhampton UK Department of Electrical Engineering Umm Al-Qura University Makkah Saudi Arabia Department of Electrical Engineering College of Engineering and Architecture Umm Al-Qura University Makkah Saudi Arabia Department of Computer Engineering College of Computer Science King Saud University Riyadh Saudi Arabia

To promote a patient-centered, sustainable healthcare ecosystem, this paper explores how blockchain and Federated Learning (FL) might be integrated into the Medical Internet of Things (MIoT). A decentralized architecture for improving data security, privacy, and interoperability between healthcare systems is proposed in this article. The article handles the drawbacks of centralized Machine Learning (ML) techniques, which frequently jeopardize the privacy of sensitive medical data, by utilizing the collaborative character of FL and the security aspects of blockchain. ML models are jointly trained while protecting patient privacy by leveraging distributed MIoT data. To effectively anticipate diseases, the study uses a variety of approaches, including FL coupled with AdaBoost, Extra Tree Classifier, Decision Tree (DT), or Linear Discriminant Analysis (LDA). Extensive testing encompassing feature selection, data splitting, cross-validation, and hyperparameter tuning guarantees the effectiveness and confidentiality of the suggested methodology. To assess the efficacy of the technique, performance measures such as Accuracy, Balanced Accuracy (BA), Fowlkes-Mallows Index (FM), Matthews Correlation Coefficient (MCC), and Bookmaker Informedness (BI) are calculated. The findings demonstrate an improved level of accuracy in contrast with conventional techniques. The research represents a novel approach to medical data analysis by choosing the best algorithm using several evaluation metrics and then incorporating it into FL.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Smartphone Application As An Air Quality Monitor Using Raspberry Pi for Reducing Air Pollution 2

Smartphone Application As An Air Quality Monitor Using Raspb...

引用

2nd International Conference on Innovative and Creative information Technology, ICITech 2021

作者： Siagian, Pandapotan Fernando, Erick Computer Engineering Study Program Del Institute of Technology Faculty of Informatics and Electrical Engineering North Sumatera22381 Indonesia School of Information Systems Bina Nusantara University Information Systems Department Jakarta11480 Indonesia

ISBN: (纸本)9781728197463

Good air quality is very important in human health. However, air pollution has been polluted everywhere, whether in the house, in cities. This is caused by a growing industry that produces air pollution. Room air pollution and various cities will contribute to air pollution, which will cause problems for human health. This work measures the carbon monoxide level using the MQ-7 sensor and measures the air's carbon dioxide level with the CCS811 sensor and MH-Z19 sensor. All three sensors are connected to the NodeMCU ESPS266, and we call it a NodeMCU module. Tests are carried out in three rooms in the house, and then three NodeMCU modules are used so that the three modules will send data alternately. Raspberry Pi3 (RPi3) functions as a server that collects data and acquires information data from sensors. Data from the CCS811 (TCOV) sensor, MH-Z19 (co2), will be processed first with fuzzy logic rules to make decisions about CO2, which are right, moderate, nasty conditions. The process is carried out on the Raspberry Pi 3 server with MQTT using publication, customer, fuzzy logic, and broker services. The value of the sensor's measurement results can be accessed via a smartphone and provides good air control values. © 2021 IEEE.

关键词： Air quality

来源：评论

学校读者我要写书评

暂无评论

Long Short Term Memory Networks for Stroke Activity Recognition base on Smartphone 5

Long Short Term Memory Networks for Stroke Activity Recognit...

引用

5th IEEE International Conference on information Technology, information systems and electrical engineering, ICITISEE 2021

作者： Siagian, Pandapotan Fernando, Erick Del Institute Of Technology Faculty Of Informatics And Electrical Engineering Computer Engineering Study Program North Sumatera 22381 Indonesia Bina Nusantara University School Of Information Systems Information Systems Department Jakarta11480 Indonesia

ISBN: (纸本)9781665401968

Aims this study to analyze rehabilitation exercises using sensor data embedded in smartphones is widely used to recognize human activities regularly to gain a better understanding of human behavior. However, it is rarely used to measure accelerators of patient movement in hospital wards. This paper proposes an optimized neural network-based feature extraction approach based on the Long Short Term Memory technique to recognize human activity using data from a tri-axial accelerometer. The activity to be measured is the Range of Motion for patient rehabilitation, and rehabilitation exercises are one of the most important steps for recovery. The results of the softmax regression trained on the dataset set have 1, 076, 105 records and 12 class attributes (class distribution). The experimental results on the SAR dataset show that our Long Short Term Memory-based approach is practical and successful, with an accuracy of 89.74%. © 2021 IEEE.

关键词： Smartphones

来源：评论

学校读者我要写书评

暂无评论

Blockchain-Empowered Lifecycle Management for AI-Generated Content (AIGC) Products in Edge Networks

TechRxiv

引用

TechRxiv 2023年

关键词： Blockchain

来源：评论

学校读者我要写书评

暂无评论

A Kolmogorov-Arnold Network for Explainable Detection of Cyberattacks on EV Chargers

arXiv

引用

arXiv 2025年

作者： Saber, Ahmad Mohammad Santos, Max Mauro Dias Janaideh, Mohammad Al Youssef, Amr Kundur, Deepa Department of Electrical and Computer Engineering University of Toronto TorontoON Canada Department of Electronics Federal Technological University of Paraná PR Ponta Grossa Brazil Department of Mechanical Engineering University of Guelph GuelphON Canada Concordia Institute for Information Systems Engineering MontrealQC Canada

The increasing adoption of Electric Vehicles (EVs) and the expansion of charging infrastructure and their reliance on communication expose Electric Vehicle Supply Equipment (EVSE) to cyberattacks. This paper presents a novel Kolmogorov-Arnold Network (KAN)-based framework for detecting cyberattacks on EV chargers using only power consumption measurements. Leveraging the KAN’s capability to model nonlinear, high-dimensional functions and its inherently interpretable architecture, the framework effectively differentiates between normal and malicious charging scenarios. The model is trained offline on a comprehensive dataset containing over 100,000 cyberattack cases generated through an experimental setup. Once trained, the KAN model can be deployed within individual chargers for real-time detection of abnormal charging behaviors indicative of cyberattacks. Our results demonstrate that the proposed KAN-based approach can accurately detect cyberattacks on EV chargers with Precision and F1-score of 99% and 92%, respectively, outperforming existing detection methods. Additionally, the proposed KANs’s enable the extraction of mathematical formulas representing KAN’s detection decisions, addressing interpretability, a key challenge in deep learning-based cybersecurity frameworks. This work marks a significant step toward building secure and explainable EV charging infrastructure. © 2025, CC BY-NC-ND.

关键词： Network security

来源：评论

学校读者我要写书评

暂无评论

DMCNET: Data-Driven Multi-Pilot Convolution Neural Network for Mimo-Ofdm Receiver 8

DMCNET: Data-Driven Multi-Pilot Convolution Neural Network f...

引用

8th IEEE International Conference on Network Intelligence and Digital Content, IC-NIDC 2023

作者： Xin, Yutong Peng, Jingyi Lu, Zejian Lee, Yong Yang, Yang China Industrial Control Systems Cyber Emergency Response Team 100029 China China Academic of Electronics and Information Technology 100041 China School of Artificial Intelligence Beijing University of Posts & Telecommunications 100876 China Collage of Electrical and Computer Engineering Sungkyunkwan University 16419 Korea Republic of

ISBN: (纸本)9798350317923

This paper focuses on studying the deep learning (DL) application of neural networks to solve the reception of single-antenna OFDM signals. Specifically, in multi-antenna scenarios, the channel model is more complex compared to single-antenna cases. By leveraging the characteristics of DL, such as automatic learning of parameters using deep neural networks, we treat the reception process of MIMOOFDM signals as a black box and utilize neural networks to accomplish the signal reception task. Moreover, we propose the data-driven multi-pilot convolution neural network for MIMO-OFDM receiver (DMCNet). By incorporating complex convolution and complex fully connected structures, we design a receiver network to recover the transmitted signals from the received signals. We validate the accuracy and robustness of DMCNet under different channel conditions, comparing the bit error rates with different schemes. Additionally, we discuss the factors influencing various channel effects. Experimental results demonstrate that the DL-based reception scheme exhibits promising feasibility in terms of accuracy and interference resistance when compared to traditional approaches. © 2023 IEEE.

关键词： MIMO 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：