检索结果-内蒙古大学图书馆

International Conference on Machine Learning and Cybernetics (ICMLC)

作者： Angga Wahyu Wibow Eri Sato-Shimokawara Kurnianingsih Department of Computer Science Graduate School of System Design Tokyo Metropolitan University Tokyo Japan Department of Electrical Engineering Politeknik Negeri Semarang Semarang Indonesia

ISBN: (数字)9798331528041

ISBN: (纸本)9798331528058

The field of Human Activity Recognition (HAR) is growing significantly in several areas but little research focuses on cultural behavior. How machine learning can explain human activity as a promotional tool in understanding cultural differences in a region is very challenging. Studies using Recurrent Neural Network (RNN), especially Long Short-Term Memory (LSTM), require large data when used to research human activity recognition. Using small data is a challenge when using LSTM. This study aims to determine which data augmentation is most appropriate when used using LSTM to predict Japanese greeting gestures, namely eshaku, keirei, saikeirei, waving hands, and Indonesian greeting gestures, namely the movement of bringing both hands together in front of the chest with a slight bow of the head. This study proposes to compare several data augmentations, namely jittering, scaling, permutation, cropping, reverse, time shifting, frequency domain, and synthetic data generation. We evaluate model performance using the Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and coefficient of determination (R 2 ) metrics. The experimental results show that Time Shifting has the best MSE, RMSE, and MAE values compared to others and Frequency Domain has the best R 2 values compared to others. This study concludes that Time Shifting and Frequency Domain are the best data augmentation methods when using LSTM in predicting greeting gestures using small data. These findings provide valuable insights into the effectiveness of various data augmentation methods in motion prediction tasks with limited data sets, as well as the importance of selecting appropriate data augmentation methods in human motion analysis. Future work could use data from two or more people using model explanations generated by machine learning and developing a new method for cultural behavior.

关键词： Hands Recurrent neural networks Frequency-domain analysis Machine learning Data augmentation Human activity recognition Cultural differences Motion analysis Long short term memory Synthetic data

来源：评论

学校读者我要写书评

暂无评论

Frequency Bandwidth Reduction of High-Frequency Intra-Body Propagation Characteristics Used for Attacker's Touch Detection in Fingerprint Authentication

Frequency Bandwidth Reduction of High-Frequency Intra-Body P...

引用

IEEE Global Conference on Consumer Electronics (GCCE)

作者： Shunichi Tomura Koki Yamada Takahiro Yoshida Department of Electrical Engineering Graduate School of Engineering Tokyo University of Science Tokyo Japan Department of Electrical Engineering and Computer Science Tokyo University of Agriculture and Technology Tokyo Japan Department of Electrical Engineering Faculty of Engineering Tokyo University of Science Tokyo Japan

In previous studies, we have proposed the attacker’s touch detection method in fingerprint authentication using high-frequency intra-body propagation characteristics to detect attacks in that an attacker holds the user's finger/wrist and presses the user's finger to a fingerprint scanner. However, the previous method used a wide range of frequency bands from 300 kHz to 10 MHz, and this caused problems in the circuit scale, cost, and size of the device in development for practical use. Therefore, we searched effective frequency bands and applied K-nearest neighbor with data augmentation in this study. As results, we achieved a 99.1 % touch detection rate with a single frequency and a 99.9 % touch detection rate with a combination of two frequencies. We achieved a significant reduction in the frequency bandwidth of the propagation characteristics.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Zero Sequence Current Detection Method of High Impedance Faults Using Distortion Waveform in Distribution Networks

Zero Sequence Current Detection Method of High Impedance Fau...

引用

IEEE PES Asia-Pacific Power and Energy engineering Conference (APPEEC)

作者： Jiraphat Manmoh Watcharin Srirattanawichaikul Department of Electrical Engineering Faculty of Engineering Chiang Mai University Chiang Mai Thailand Graduate Program in Electrical Engineering Graduate School Chiang Mai University Chiang Mai Thailand Provincial Electricity Authority Bangkok Thailand

ISBN: (数字)9798350318098

ISBN: (纸本)9798350318104

This paper presents high-impedance faults (HIF) detected in distribution networks with distortion zero sequence current detection methods. The simulation model in this paper is performed on a single branch distribution network with a voltage rating of 22kV, and the simulation is divided into two parts. The first part is the creation of HIF in the system by adjusting the distortion of the HIF current according to the principle of the heat-balanced equation of the material. The second part takes the resulting HIF model, which feeds into the established distribution network and detects distortion zero sequence currents occurring. The simulation results clearly depict the current waveform distortion caused by HIF, which can be adjusted. The presence of zero sequence current can indicate the occurrence of HIF in the distribution network.

关键词： Heating systems Surface impedance Surface waves Simulation Distribution networks Voltage Distortion

来源：评论

学校读者我要写书评

暂无评论

Greeting Gesture Analysis Using Boosting Algorithms and SHAP

Greeting Gesture Analysis Using Boosting Algorithms and SHAP

引用

International Symposium on Soft Computing and Intelligent Systems (SCIS)

作者： Angga Wahyu Wibowo Eri Sato-Shimokawara Kurnianingsih Department of Computer Science Graduate School of Systems Design Tokyo Metropolitan University Tokyo Japan Department of Electrical Engineering Politeknik Negeri Semarang Semarang Indonesia

ISBN: (数字)9798350373332

ISBN: (纸本)9798350373349

Recently, studies related to human activity recognition have developed which have been applied in various fields. In that field Machine learning and deep learning techniques have been widely used for several classification and prediction tasks. However, the first problem is that there is rarely any research on human activities that reflect the culture of a region, such as greeting gestures. Second, there are few studies that do not provide models resulting from the classification or prediction process using machine learning, which can result in a lack of understanding of how a model from machine learning is generated. This study aims to predict Japanese greeting gestures (eshaku, keirei, saikeirei, waving hand) and Indonesian (placing both hands in front of the chest) using Light Gradient Boosting Machine (LightGBM), eXtreme Gradient Boosting (XGBoost), and Gradient Boosting. Then, represent the model of the learning machine using SHapley Additive exPlanations (SHAP) to find out how the learning machine obtain a model. Our findings show the efficacy of this method and show that the model not only predicts greeting gestures with high accuracy but also offers a transparent understanding of the factors influencing its predictions, thus enhancing the interpretability and applicability of machine learning in recognizing culturally significant human activities. In the future, we plan to use machine learning and other Explainable Artificial Intelligence (XAI) methods to study human activities with cultural concepts such as greeting gestures in different styles, create more greeting gesture-based human activity datasets, and examine the ethical implications or considerations of using machine learning in cultural contexts to respect cultural differences.

关键词： Deep learning Ethics Machine learning algorithms Additives Explainable AI Machine learning Predictive models Boosting Cultural differences Human activity recognition

来源：评论

学校读者我要写书评

暂无评论

E-Payment System to Reduce Use of Paper Money for Daily Transactions 4

E-Payment System to Reduce Use of Paper Money for Daily Tran...

引用

4th International Conference on electrical, computer and Telecommunication engineering, ICECTE 2022

作者： Tayan, Omar Salem, Golam Saleh Ahmed Alginahi, Yasser M Kabir, Muhammad Nomani Department of Scientific Research and Graduate Studies Al-Madinah Al-Munawarrah Saudi Arabia Trust University Department of Electrical and Electronics Engineering Barishal8200 Bangladesh Adrian College Department of Computer Science AdrianMI United States Trust University Department of Computer Science and Engineering Barishal8200 Bangladesh

ISBN: (纸本)9798350320541

This paper presents a framework of an E-payment system which can be used for daily petty cash transactions. The framework involves communication of encrypted messages through an insecure network using Advanced Encryption System (AES) with a timestamp key. To ensure the message authentication and integrity, the Hash-based Message Authentication Code (HMAC) is used. The associated key in HMAC is considered as a function of the phone Media Access Control (MAC) address as well as the user's biometric data for additional security. Timestamp key is updated each time a transaction takes place. Thus, the framework can be implemented to replace the use of paper money for daily transactions. © 2022 IEEE.

关键词： Cryptography

来源：评论

学校读者我要写书评

暂无评论

Light Shift Detection Using a Multiple Photodetection Method Suitable for a Chip-Scale Atomic Clock

Light Shift Detection Using a Multiple Photodetection Method...

引用

IEEE International Frequency Control Symposium

作者： Masahiro Fukuoka Shigeyoshi Goka Department of Electrical Engineering and Computer Science Graduate School of System Design Tokyo Metropolitan University Tokyo Japan

SummaryIn this study, we demonstrated light shift detection in a running atomic clock using the multi-photodetection method. This method enables the observation of atomic resonances with different intensities using the intensity distribution of the laser beam. As it requires only two photodiodes to detect light shifts, it is easily applicable to miniature vapor-cell-based atomic clocks such as chip-scale atomic clocks and miniature atomic clocks.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Overcoming Data Inequality across Domains with Semi-Supervised Domain Generalization

arXiv

引用

arXiv 2024年

作者： Park, Jinha Cho, Wonguk Kim, Taesup Department of Electrical and Computer Engineering Seoul National University Korea Republic of Graduate School of Data Science Seoul National University Korea Republic of

While there have been considerable advancements in machine learning driven by extensive datasets, a significant disparity still persists in the availability of data across various sources and populations. This inequality across domains poses challenges in modeling for those with limited data, which can lead to profound practical and ethical concerns. In this paper, we address a representative case of data inequality problem across domains termed Semi-Supervised Domain Generalization (SSDG), in which only one domain is labeled while the rest are unlabeled. We propose a novel algorithm, ProUD, which can effectively learn domain-invariant features via domain-aware prototypes along with progressive generalization via uncertainty-adaptive mixing of labeled and unlabeled domains. Our experiments on three different benchmark datasets demonstrate the effectiveness of ProUD, outperforming all baseline models including single domain generalization and semi-supervised learning. Source code will be released upon acceptance of the paper. Copyright © 2024, The Authors. All rights reserved.

关键词： Population statistics

来源：评论

学校读者我要写书评

暂无评论

FCS-MPC Based Dual-module ANN Controller for Three-level Converter

FCS-MPC Based Dual-module ANN Controller for Three-level Con...

引用

International Conference on Power Electronics (ICPE)

作者： Kun Wang Xinliang Yang Shuiqi Chen Ki-Bum Park Cho Chun Shik Graduate School of Mobility KAIST Korea Electrical and Computer Engineering Department UCLA USA

Finite-control-set model predictive control (FCS-MPC) shows great control performance and adaptability for different converter topologies and operating modes. However. The computation burden increases significantly for long prediction step and multi-level topology. Artificial neuron network (ANN) is developed to imitate FCS-MPC controller for similar control effect with lower computation burden. However, the imitation accuracy is not good enough for single ANN. To achieve acceptable control effect using a simple ANN, we propose an FCS-MPC-based dual-module ANN controller. We first off-line train the ANN to imitating the FCS-MPC. Then we designed a dual-module structure which combines ANN and FCS-MPC to increase the imitation accuracy. The simulation result shows that the accuracy of our design increases to 99.87% while the computation burden is reduced by 58.8% compared with FCS-MPC. It can achieve similarly control performance and significantly reduce computation burden.

关键词：

来源：评论

学校读者我要写书评

暂无评论

High-Resolution Hydrogen Molecule Detection Utilizing a Nanocandle-Based Optical Sensor 16

High-Resolution Hydrogen Molecule Detection Utilizing a Nano...

引用

16th Pacific Rim Conference on Lasers and Electro-Optics, CLEO-PR 2024

作者： Park, Nu-Ri Kim, Min-Joong Lee, Jongsu Seo, Minah Ryu, Yong-Sang Kim, Myung-Ki KU-KIST Graduate School of Converging Science and Technology Korea University Seoul02841 Korea Republic of Korea Institute of Science and Technology Sensor System Research Center Seoul02792 Korea Republic of Northwestern University Department of Electrical Engineering and Computer Science EvanstonIL60208 United States School of Biomedical Engineering Korea University Seoul02481 Korea Republic of Interdisciplinary Program in Precision Public Health Korea University Seoul02481 Korea Republic of

Our research introduces a novel method for safe hydrogen detection. We've developed an advanced nano-candle sensor, combining Pd with nano-candles, enabling accurate detection of low-concentration hydrogen ( ©... 详细信息

ISBN: (纸本)9798350372076

关键词： Nanoclay

来源：评论

学校读者我要写书评

暂无评论

Classification Methods for Recognition of Power Quality Disturbances in Distribution Networks Using Artificial Neural Network

Classification Methods for Recognition of Power Quality Dist...

引用

IEEE PES Asia-Pacific Power and Energy engineering Conference (APPEEC)

作者： Wassanan Buasi Watcharin Srirattanawichaikul Department of Electrical Engineering Faculty of Engineering Chiang Mai University Chiang Mai Thailand Graduate Program in Electrical Engineering Graduate School Chiang Mai University Chiang Mai Thailand Provincial Electricity Authority Bangkok Thailand

ISBN: (数字)9798350318098

ISBN: (纸本)9798350318104

This paper provides the power quality disturbances (PQDs) classification in distribution networks using an artificial neural network (ANN). Various forms of PQDs are considered, including sag, swell, harmonics, flickers, transients, sag with harmonics, and swell with harmonics. The simulation is divided into two parts. The first part is PQDs signals creation and feature extraction using discrete wavelet transform (DWT) and classification by using the dataset from the feature extraction step to be the input of artificial neural network (ANN). The second part is the model testing with the IEEE 9 bus test system with 5 PQDs, which is the most misclassified. The results of the proposed method will be demonstrated using simulation results in MATLAB/Simulink. The results show that the model can classify PQD signals into ten classes. The accuracy of the model is at 99.4%. After testing with the IEEE 9 bus test system, the total accuracy is at 95.6%.

关键词： Accuracy Power quality Artificial neural networks Distribution networks Feature extraction Harmonic analysis Mathematical models

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：