检索结果-内蒙古大学图书馆

IEEE International Conference on Cognitive Machine Intelligence (CogMI)

作者： Dohee Kim Sunghyun Sim Bori Yoon Ling Liu Hyerim Bae Dept. Industrial Engineering Major in Industrial Data Science & Engineering Pusan National University Busan South Korea Dept. Industrial Management & Big Data Engineering Dong-Eui University Busan South Korea College of Computing Georgia Institute of Technology Georgia USA

Time-series data consists of a sequence of observations recorded in chronological order, where the data changes over time. This type of data exhibits various characteristics, such as temporal volatility, trends, and seasonality. Recently, a new layer structure called Correlation Recurrent Units (CRU) has been proposed to capture not only temporal variability but also trends and seasonality in time-series data. In this study, we propose an end-to-end model that utilizes a CRU autoencoder to learn temporal feature representations and address time-series classification problems simultaneously. To validate the performance of our proposed model, we conducted comparative experiments using 30 time-series classification datasets from six different types. The experimental results showed that the proposed model outperformed the baseline models in 20 out of the 30 time-series datasets. This indicates that the proposed approach effectively captures various temporal features in time-series data and improves the performance of time-series classification tasks.

关键词：

来源：评论

学校读者我要写书评

暂无评论

CAD-DA: Controllable Anomaly Detection after Domain Adaptation by Statistical Inference 27

CAD-DA: Controllable Anomaly Detection after Domain Adaptati...

引用

27th International Conference on Artificial Intelligence and Statistics, AISTATS 2024

作者： Le Duy, Vo Nguyen Lin, Hsuan-Tien Takeuchi, Ichiro University of Information Technology Ho Chi Minh City Viet Nam Vietnam National University Ho Chi Minh City Viet Nam Dept. of Computer Science and Information Engineering National Taiwan University Taiwan Center for Data Intelligence National Taiwan University Taiwan Nagoya University Japan RIKEN Japan

We propose a novel statistical method for testing the results of anomaly detection (AD) under domain adaptation (DA), which we call CAD-DA—controllable AD under DA. The distinct advantage of the CAD-DA lies in its ability to control the probability of misidentifying anomalies under a prespecified level α (e.g., 0.05). The challenge within this DA setting is the necessity to account for the influence of DA to ensure the validity of the inference results. We overcome the challenge by leveraging the concept of Selective Inference to handle the impact of DA. To our knowledge, this is the first work capable of conducting a valid statistical inference within the context of DA. We evaluate the performance of the CAD-DA method on both synthetic and real-world datasets. Copyright 2024 by the author(s).

关键词： Anomaly detection

来源：评论

学校读者我要写书评

暂无评论

Deep Learning-Driven Black Box Approach For Image Segmentation

Deep Learning-Driven Black Box Approach For Image Segmentati...

引用

Advances in Computing, Communication and Networking (ICAC2N), International Conference on

作者： Nuthalapati Sudha Sai Bhargav Kasetty K. Sai Madhuri Jajala Nikitha Issac Neha Margret K. Rajakumar School of Engineering Mahindra University Hyderabad Telangana India School of Computer Science and Engineering V.I.T – University Vellore Tamilnadu India Dept. of Computer Science and Data Science BVRIT Narsapur Hyderabad Telangana India Dept. of CSE – AIML & IoT VNR VIET Hyderabad Telangana India

ISBN: (数字)9798350356816

ISBN: (纸本)9798350356823

The diagnosis of a range of eye disorders needs to categorize the retinal vessels. Computerized implementation of this process is becoming increasingly essential for automated screening systems for retinal diseases. To achieve a more accurate extraction of the retinal vessels, a new pre-processing step is proposed. These proposed pre-processes are also compared to other algorithms to assess their impact. The proposed pre-processing process consists of two phases. The first phase is the implementation and validation of the pre-processing modules, and the second phase is the implementation of these pre-processes onto the retinal vessels that were to be extracted. To achieve a significantly improved segmented vessel image, the proposed pre-process phase employs a common image-processing technique. In recent years, there has been a great deal of focus on retinal vessel identification studies, and the importance of assessing and confirming the findings of retinal vessel segmentation.

关键词： Image segmentation Analytical models Accuracy Computational modeling Retinal vessels Diseases

来源：评论

学校读者我要写书评

暂无评论

Driver Behavior Identification with Facial Recognition and Deep Learning

Driver Behavior Identification with Facial Recognition and D...

引用

Communication, Computing and Energy Efficient Technologies (I3CEET), International Conference on

作者： Ajay Sudhir Bale T N Siva Kumar Keerthana Lal Arulananth T S Naveen Ghorpade Sinchana. K. P Dept. of ECE New Horizzon College of Engineering Bengaluru India Dept. of Cyber Security & Data Science Bapatla Engineering College Autonomous Bapatla Andhra Pradesh India Dept. of ECE Amrita School of Engineering Bengaluru India Department of Electronics and Communication Engineering MLR Institute of Technology Hyderabad Telangana India Dept. of CSE Vijaya Vittala Institute of Technology Bengaluru India Dept. of CSE Don Bosco Institute of Technology Bengaluru India

ISBN: (数字)9798331541583

ISBN: (纸本)9798331541590

Driving entails a series of actions demanding intense focus. Some- times, these actions get overshadowed by other behaviors such as enjoy engaging in ingesting activities, conversing, using a cellphone, tuning the radio, or even drowsiness. These behaviors stand as the primary culprits behind today's traffic accidents. Thus, it's imperative to create applications that can forewarn drivers. This study proposes a streamlined convolutional neural network structure for identifying driver behaviors, thereby enhancing the accuracy of the alert system and minimizing traffic collisions. This network combines modules for extracting features and making classifications. By merging these components, the feature extraction within the neural network is designed to be more potent and efficient. The model aims to decrease computational costs and amplify data flow within the network by employing dept.-wise separable convolutions. Additionally, the Convolution Block Attention Module aids the network in prioritizing vital features, resulting in improved performance for various computer vision tasks. The classifier unit includes global average pooling and a softmax layer to compute class probabilities. The overall architecture optimizes network parameters while upholding classification precision. Three standard datasets are used for both training and assessing the whole network.

关键词： Training Accuracy Computational modeling Feature extraction Real-time systems Telecommunication computing Convolutional neural networks Driver behavior Tuning Videos

来源：评论

学校读者我要写书评

暂无评论

Heart Abnormality Detection Through Neural Network

Heart Abnormality Detection Through Neural Network

引用

Self Sustainable Artificial Intelligence Systems (ICSSAS), International Conference on

作者： Sarada B Smitha Chowdary Ch Jampani Satish Babu K. Kavitha Dept. of Data Science Malla Reddy Engineering College for Women(JNTUH) Secunderabad India Dept. of CSE Koneru Lakshmiah Education Foundation Guntur India Dept. of Computer Science P.B.Siddhartha College of Arts and Science Vijayawada India

This research study aims to develop an algorithm using phonocardiograms and Deep learning to differentiate between normal and abnormal heart sounds. The increase in cardiovascular diseases (CVD's), primarily caused by unhealthy habits, necessitates efficient diagnosis. The study utilized 1055 data samples, with 833 being normal and 182 abnormal, each lasting 5 seconds. Logistic regression and a three-layered neural network were initially applied, yielding accuracies of 42% and 63%, respectively. By extracting features using spectral roll-off and reapplying the models, accuracies improved to 79% and 86%. Up sampling the data further increased the accuracies to 86% for logistic regression and 89% for the three-layered neural network. The findings demonstrate the potential of audio-based algorithms in diagnosing cardiovascular diseases accurately.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Ocular Cataract Identification Using Deep Convolutional Neural Networks

Ocular Cataract Identification Using Deep Convolutional Neur...

引用

International Conference on Advances in Big data, Computing and data Communication Systems (icABCD)

作者： Feliciana M. E. Manuel Saide Manuel Saide Felermino D. M. A. Ali Sanae Lotfi Dept. of Computer Engineering Faculty of Engineering Lurio University Pemba Mozambique Center for Data Science New York University New York USA

Ocular cataract is among diseases that result in blindness if not treated in time. It affects people worldwide, primarily in underdeveloped countries. This health problem affects the quality of patients' lives. However, early diagnosis avoids blindness and allows the patient to have appropriate treatment. Developing countries, especially those with low income, have a precarious health system, even in the ophthalmology sector, where equipment is lacking. This research aims to develop a deep learning-based model to detect ocular cataracts based on retinal images. We collect 1000 retinal images from Kaggle, which are then equally divided into two classes: with and without cataracts. We then use several neural architectures to correctly classify these images, including ResNet18, ResNet34, InceptionResNetV2, and InceptionV4. We demonstrate that ResNet18 outperforms the other architectures, reaching 95.5% accuracy score. Our results suggest that deep convolutional neural networks can achieve a significant performance in ocular cataracts classification using retinal images.

关键词：

来源：评论

学校读者我要写书评

暂无评论

The Impact of the Energy Storage System Operation Strategy on Power Grid Reliability

引用

Transactions of the Korean Institute of Electrical Engineers 2021年第12期70卷 1820-1826页

作者： Kim, Beopsoo Lee, Dayoung Rusetskii, Nikita Shusterzon, Konstantin Sidorov, Denis Kim, Insu Dept. of Electrical Engineering Inha University Korea Republic of Dept. of Electrical and Computer Engineering Inha University Korea Republic of School of Information Technology and Data Science Irkutsk National Research Technical University Russia Energy Systems Institute Siberian Branch Russian Academy of Sciences Russia

Based on the empirical data published by the Korea Electric Power Exchange and the Korea Meteorological Administration, we present a methodology that predicts the system marginal prices. For this purpose, this study uses the long-short term memory (LSTM) technique, which is commonly used in time-series prediction problems. As a result, we present a strategy for operating energy storage systems (ESSs) from the perspective of a prosumer rather than a system operator or conventional utilities. This study also examines the effects of the proposed prosumer actions on the system. The proposed operating strategy is established based on dynamic programming. In particular, this study takes the proposed operation strategy of ESSs the point of view of system reliability. © The Korean Institute of Electrical Engineers This is an Open-Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://***/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

关键词： Reliability

来源：评论

学校读者我要写书评

暂无评论

Summarization of Investment Reports Using Pre-trained Model 14

Summarization of Investment Reports Using Pre-trained Model

引用

14th IIAI International Congress on Advanced Applied Informatics, IIAI-AAI 2023

作者： Sakaji, Hiroki Kobayashi, Ryotaro Izumi, Kiyoshi Mitsugi, Hiroyuki Kuramoto, Wataru The University of Tokyo School of Engineering Tokyo Japan Daiwa Institute of Research Ltd. Frontier Technologies Research Center. Data-driven Science Dept. Tokyo Japan Daiwa Asset Management Co.Ltd. Fund Management Division Tokyo Japan

ISBN: (纸本)9798350324228

In this paper, we attempt to summarize monthly reports as investment reports. Fund managers have a wide range of tasks, one of which is the preparation of investment reports. In addition to preparing monthly reports on fund management, fund managers prepare management reports that summarize these monthly reports every six months or once a year. The preparation of fund reports is a labor-intensive and time-consuming task. Therefore, in this paper, we tackle investment summarization from monthly reports using transformer-based models. There are two main types of summarization methods: extractive summarization and abstractive summarization, and this study constructs both methods and examines which is more useful in summarizing investment reports. © 2023 IEEE.

关键词： Investments

来源：评论

学校读者我要写书评

暂无评论

Neural Network Coding in data Compression Systems in Communication Channels 6

Neural Network Coding in Data Compression Systems in Communi...

引用

6th International Scientific Conference on Information, Control, and Communication Technologies, ICCT 2022

作者： Berezkin, Aleksandr Kukunin, Dmitry Kirichek, Ruslan State University of Telecommunications Bonch-Bruevich Saint-Petersburg Dept. of Software Engineering and Computer Science St. Petersburg Russia Bonch-Bruevich Saint-Petersburg State University of Telecommunications Dept. of Communication Networks and Data Transmission St. Petersburg Russia

ISBN: (数字)9781665493673

ISBN: (纸本)9781665493673

The article describes the use of a neural network implementation of BCH codes encoding and decoding devices for error correction in on-the-fly data compression systems. The user payload vector is encoded in the BCH code with additional redundancy, which provides linear data separability and better convergence when training the autoencoder. On the receiving side, the decoding device operates in the error correction mode, which makes it possible not to impose strict requirements on the accuracy of data recovery when training the autoencoder. Examples of such devices are described in the article. © 2022 IEEE.

关键词： data compression

来源：评论

学校读者我要写书评

暂无评论

Blockchain-Based Music Streaming Services: A Novel Approach to Transparent Artist Control

Blockchain-Based Music Streaming Services: A Novel Approach ...

引用

Blockchain and Distributed Systems Security (ICBDS), IEEE International Conference on

作者： Rashmi Ashtagi Tanmay Deshmukh Swati Jadhav Ranjeet Bidwe Afsha Akkalkot Amar Buchade Nilesh Uke Sudipta Banerjee Dept. of Computer Engineering Vishwakarma Institute of Technology Pune India Dept. of CSE Symbiosis Institute of Technology Pune India Dept. Name of Computer Engineering COEP Technological University Pune India Dept. of Artificial Intelligence Data Science Vishwakarma Institute Technology Pune India Indira College of Engineering and Management Pune India Symbiosis Institute of Technology Pune India

ISBN: (数字)9798350354348

ISBN: (纸本)9798350354355

Artists have historically used music as a form of personal expression. However, due to the centralized nature of music streaming services, artists receive only a small percentage of the revenue generated and have no control over their content produced. To address these issues, a revolutionary technique is needed that provides artists with greater control and transparency over their output. This paper focuses on a decentralized strategy that utilizes blockchain technology to facilitate direct interaction between artists and users, enabling artists to upload their music, sell songs, and receive royalties. The implementation of blockchain technology could potentially revolutionize the music industry by ensuring that rightful owners receive payment for their music while reducing the waste of paying intermediaries. Without intermediaries involved in the distribution process, the music industry could pay out more royalties to rights holders automatically and instantaneously. The proposed solution also addresses piracy concerns through the use of encryption and watermarking techniques. Blockchains provide security, confidentiality, tamper-proofing, traceability, control, and oversight over information. By developing a secure, self-installing, cross-platform reader positioned on the header of the digital content file, The research empowers data owners to manage the flow of information even after it has been made available. The suggested approach enables content owners to control their digital content and monitor its most popular components after distribution. The blockchain-based system also benefits from token distribution among the content’s owner, distributor, and consumer. The simulation results demonstrate that our approach is capable of overcoming the flaws of the existing system.

关键词： Industries Access control Simulation Smart contracts Process control Watermarking Blockchains Encryption Monitoring

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：