检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Xu, Moucheng Zhou, Yukun Goodwin-Allcock, Tobias Firoozabadi, Kimia Jacob, Joseph Alexander, Daniel C. Slator, Paddy J. UCL Hawkes Institute United Kingdom UCL Satsuma Lab United Kingdom UCL Respiratory United Kingdom UCL Department of Computer Science United Kingdom UCL Medical School United Kingdom UCL Department of Medical Physics and Biomedical Engineering United Kingdom Cardiff University School of Computer Science and Informatics United Kingdom Cardiff University Brain Research Imaging Center United Kingdom

We introduce and demonstrate a new paradigm for quantitative parameter mapping in MRI. Parameter mapping techniques, such as diffusion MRI and quantitative MRI, have the potential to robustly and repeatably measure biologically-relevant tissue maps that strongly relate to underlying microstructure. Quantitative maps are calculated by fitting a model to multiple images, e.g. with least-squares or machine learning. However, the overwhelming majority of model fitting techniques assume that each voxel is independent, ignoring any co-dependencies in the data. This makes model fitting sensitive to voxelwise measurement noise, hampering reliability and repeatability. We propose a self-supervised deep variational approach that breaks the assumption of independent pixels, leveraging redundancies in the data to effectively perform data-driven regularisation of quantitative maps. We demonstrate that our approach outperforms current model fitting techniques in dMRI simulations and real data. Especially with a Gaussian mixture prior, our model enables sharper quantitative maps, revealing finer anatomical details that are not presented in the baselines. Our approach can hence support the clinical adoption of parameter mapping methods such as dMRI and qMRI. Our code is available at https://***/moucheng2017/MRI-GMM-VAE. © 2024, CC BY-NC-ND.

关键词： Mapping

来源：评论

学校读者我要写书评

暂无评论

Enhancing Sentiment Classification in Twitter Data Through Context-Driven Text Processing and Tweet Embeddings

Enhancing Sentiment Classification in Twitter Data Through C...

引用

IEEE Region 10 Humanitarian Technology Conference (R10-HTC)

作者： Vassilis C. Gerogiannis Andreas Kanavos Nikos Antonopoulos Amrita Bhola Biswaranjan Acharya Department of Digital Systems University of Thessaly Larissa Greece Department of Informatics Ionian University Corfu Greece Department of Digital Media and Communication Ionian University Argostoli Greece Indira Gandhi Institute of Technology Sarang Odisha India Department of Computer Engineering- AI & BDA Marwadi University Rajkot India

Sentiment analysis and text classification tasks heav-ily rely on text processing techniques. However, existing approaches often neglect domain-specific factors and rely on generic routines and pre-built dictionaries. In this paper, we investigate the impact of text processing steps on sentiment classification using Twitter data. Our approach introduces skip gram-based word embeddings that effectively capture Twitter-specific fea-tures, such as informal language and emojis. Through rigorous experimentation, we identify the detrimental consequences of conventional text processing steps like stop word removal and simple averaging of term vectors for tweet representation. To optimize sentiment classification, we propose new effective steps, including the inclusion of emoji characters, measuring word importance from embeddings, aggregating term vectors into tweet embeddings, and creating a linearly separable feature space. Our results demonstrate the superiority of context-driven word embeddings in selecting important words for tweet clas-sification, outperforming pre-built word dictionaries. Moreover, the proposed tweet embedding reduces reliance on multiple text processing steps, resulting in more accurate sentiment analysis on Twitter data.

关键词：

来源：评论

学校读者我要写书评

暂无评论

TIPS: A text interaction evaluation metric for learning model interpretation

引用

Expert Systems with Applications 2025年 287卷

作者： Nie, Zhenyu Xiao, Zheng Wang, Tao Chronopoulos, Anthony Theodore Andonie, Răzvan Mosavi, Amir College of Information Science and Engineering Hunan University Changsha China Department of Computer Science University of Texas San AntonioTX78249 United States Department of Computer Science Central Washington University EllensburgWA98926 United States Department of Electronics and Computers Transilvania University of Braşov Braşov500036 Romania John von Neumann Faculty of Informatics Obuda University Budapest1034 Hungary Institute of the Information Society Ludovika University of Public Service Budapest Hungary Department of Information and Computing Systems Abylkas Saginov Karaganda Technical University Karaganda Kazakhstan

Explaining the decision-making behavior of deep neural networks (DNNs) can increase their trustworthiness in real-world applications. For natural language processing (NLP) tasks, many existing interpretation methods split the text according to the interactions between words. Also, the evaluation of explanation capability focuses on justifying the importance of the divided text spans from the perspective of interaction contribution. However, the prior evaluations are misled by extra interactions, making the evaluation unable to acquire accurate interactions within the text spans. Besides, existing research considers only absolute interaction contribution, which causes the evaluation to underestimate the important text spans with lower absolute interaction contribution and to overestimate the unimportant text spans with higher absolute interaction contribution. In this work, we propose a metric called Text Interaction Proportional Score (TIPS) to evaluate faithful interpretation methods. More specifically, we use a pick scheme to acquire the interactions within the divided text span and eliminate the influence of the extra interactions. Meanwhile, we utilize the relative interaction contribution between the divided text span and whole text to measure the importance of the acquired interactions. The proposed metric is validated using two interpretation methods in explaining three neural text classifiers (LSTM, CNN and BERT) on six benchmark datasets. Experiments show that TIPS outperforms a baseline method in three ways consistently and significantly (i.e., acquiring interactions within the text span, measuring importance of interaction, and distinguishing the important and unimportant text spans). © 2025 Elsevier Ltd

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Cutting-Edge Predictive Methods for Enhancing Short-Term Load Prediction

Cutting-Edge Predictive Methods for Enhancing Short-Term Loa...

引用

2024 International Conference on Electrical and computer engineering Researches, ICECER 2024

作者： Ghazal, Taher M. Janjua, Jamshaid Iqbal Abushiba, Walid Ahmad, Munir Abbas, Tahir Ateeq, Karamath College of Arts & Science Applied Science University P.O.Box 5055 Manama Bahrain Universiti Kebangsaan Malaysia Center for Cyber Security Faculty of Information Science and Technology Bangi Selangor Malaysia Lahore Pakistan College of Engineering Applied Science University Bahrain College of Informatics Korea University Seoul02841 Korea Republic of TIMES Institute Department of Computer Science Multan Pakistan School of Computing Skyline University College University City Sharjah Sharjah United Arab Emirates

ISBN: (纸本)9798331539733

Efficient management and cost-related factors in the power sector call for accurate short-term load forecasting as it enables better planning with the electric grid and achieving stability within it. This paper looks into the state-of-the-art forecasting strategies that enhance load forecasts made in shorter time frames. Such techniques include but are not limited to machine learning applications, effective data models, and multi-dimensional modeling leveraging historical, meteorological, and economic data. Their review is carried out in detail with their advantages and disadvantages and their application in practice. Furthermore, this study strives to understand the barriers that need to be overcome in the quest to increase the accuracy of the forecasting models especially concerning renewable energy sources. We illustrate how new trends relate to future developments of predictive models about supporting the decision-making in operating power systems and improving the reliability and sustainability of energy systems. © 2024 IEEE.

关键词： Prediction models

来源：评论

学校读者我要写书评

暂无评论

Comparison of Smoothing Methods to Remove Artifacts in Emotion Recognition based on Electroencephalogram Signals

Comparison of Smoothing Methods to Remove Artifacts in Emoti...

引用

International Conference on informatics and Computing (ICIC)

作者： I Made Agus Wirawan Retantyo Wardoyo Danang Lelono Sri Kusrohmaniah Education of Informatics Engineering Department Universitas Pendidikan Ganesha Bali Indonesia Department of Computer Science and Electronics Universitas Gadjah Mada Yogyakarta Indonesia Department of Psychology Universitas Gadjah Mada Yogyakarta Indonesia

ISBN: (纸本)9798350345728

Although the baseline reduction approach improves emotion recognition accuracy, artifacts in the baseline Electroencephalogram signal can affect emotion recognition accuracy. This study examines and compares three smoothing methods to remove artifacts in the baseline Electroencephalogram signal to overcome these problems. Based on the experimental scenario design, the Mean filter method is proven to remove the artifacts contained in the baseline Electroencephalogram signal. This effort has an impact on increasing the accuracy of emotion recognition. Through the tests carried out on the DEAP dataset, the baseline Electroencephalogram signal feature that has been smoothed using the Mean filter method can produce a high accuracy of emotion recognition when applied to the Relative Difference method, with an average accuracy of 95.64%, 95.18%, and 92.82% for each emotion of High/Low Arousal, High/Low Valence, and four class emotion. The same thing also happened to the Fractional Difference method, where the value of the baseline Electroencephalogram signal feature that had been smoothed using the Mean filter method also produced the highest average accuracy values of 95.68%, 95.02% and 92.70% for each High/Low Arousal, High/Low Valence, and four class emotions, respectively. Based on the results of this experiment, using the baseline Electroencephalogram signal feature that has been smoothed using the Mean filter method can improve the baseline reduction approach.

关键词： Emotion recognition Smoothing methods informatics

来源：评论

学校读者我要写书评

暂无评论

Smart watering system based on framework of low-bandwidth distributed applications (LBDA) in cloud computing 12th

Smart watering system based on framework of low-bandwidth di...

引用

12th International Conference on Intelligent Networking and Collaborative Systems, INCoS 2020

作者： Sirimorok, Nurdiansyah As, Mansur Yoshida, Kaori Köppen, Mario Graduate School of Computer Science and Systems Engineering Kyushu Institute of Technology Fukuoka Japan Graduate School of Life Science and Systems Engineering Kyushu Institute of Technology Kitakyushu Japan Department of Life Science and Systems Engineering Kyushu Institute of Technology Kitakyushu Japan Department of Creative Informatics Kyushu Institute of Technology Fukuoka Japan

ISBN: (纸本)9783030577957

The estimation of soil moisture content is required for agriculture, mainly to build the irrigation scheduling model. In this study, we present a smart watering system to deal with various factors derived from the stochastic information in the agricultural operational, i.e., air temperature, air humidity, soil moisture, soil temperature, and light intensity. The methodology consists of exploiting the Internet of Things (IoT) using Low-Bandwidth Distributed Applications (LBDA) in cloud computing to integrate the real data sets collected by the several sensor technologies. We conducted experiments for the watering system used two types of soil and different plant. Here, the Long Short Term Memory Networks (LSTMs) approach in deep learning techniques used to build smart decisions concerning watering requirements and deal with heterogeneous information coming from agricultural environments. Results show that our models can effectively improve the prediction accuracy for the watering system over various soil and plants. © Springer Nature Switzerland AG 2021.

关键词： Cloud computing

来源：评论

学校读者我要写书评

暂无评论

A Hybrid Cross-Scale Transformer Architecture for Robust Medical Image Segmentation

A Hybrid Cross-Scale Transformer Architecture for Robust Med...

引用

2022 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2022

作者： Ning, Yang Zhang, Shouyi Zhong, Wei Liu, Peide Zhang, Caiming Shandong Jianzhu University School of Computer Science and Technology Jinan China The University of Edinburgh School of Informatics Edinburgh United Kingdom Guangdong Police College Department of Network Information Security Guangzhou China Shandong University of Finance and Economics School of Management Science and Engineering Jinan China Shandong University School of Software Jinan China

ISBN: (纸本)9781665468190

Transformer architecture has emerged to be successful in many natural language processing tasks. However, its applications to clinical practice remain largely unexplored. In this study, we propose a Robust Cross-Scale Hybrid Transformer (RCSHT) architecture for medical image segmentation, which can effectively enhance the multi-scale feature representations while integrating local features with global dependencies. Specifically, we propose two new modules based on the self-attention mechanism: PSCM and PCCM, which perform matrix low-rank transformation on cross-scale images in spatial and channel space, respectively, to effectively enhance the discriminant ability of long-range dependencies of multi-scale images and effectively reduce the computational complexity. Meanwhile, we use spatial positional encoding to perform spatial regularization on the two attention modules, so that the model has more opportunities to obtain valuable spatial details and further enhance the spatial retention ability. Besides, by fusing outputs of the two modules, the robustness and discriminant ability of the feature representations are further improved. Note that our hybrid architecture allows transformers to be initialized as convolutional networks without pre-training. Extensive experiments conducted on our newly proposed dataset have demonstrated our RCSHT notably outperforms the state-of-the-art methods by a large margin, holding the promise to generalize well on other medical image segmentation. © 2022 IEEE.

关键词： Signal encoding

来源：评论

学校读者我要写书评

暂无评论

ResNet-Attention model for human authentication using ECG signals

ResNet-Attention model for human authentication using ECG si...

引用

作者： Hammad, Mohamed Plawiak, Pawel Wang, Kuanquan Acharya, Udyavara Rajendra School of Computer Science and Technology Harbin Institute of Technology Harbin China Faculty of Computers and Information Menoufia University Menoufia Egypt Department of Information and Communications Technology Faculty of Computer Science and Telecommunications Cracow University of Technology Krakow Poland Institute of Theoretical and Applied Informatics Polish Academy of Sciences Gliwice Poland Department of Electronics and Computer Engineering Ngee Ann Polytechnic Singapore Singapore Department of Biomedical Engineering School of Science and Technology Singapore School of Social Sciences Singapore Singapore Department of Bioinformatics and Medical Engineering Asia University Taichung Taiwan

Authentication is the process of verifying the claimed identity of the user. Recently, traditional authentication methods such as passwords, tokens, and so on are no longer used for authentication as they are more prone to theft and different types of violations. Therefore, new authentication approaches based on biometric modalities such as heartbeat pattern obtained from electrocardiogram (ECG) signals are considered. Unlike other biometrics, ECG provides the assurance that the person is alive, and is considered as one of the most accurate recent methods for authentication. In this article, two end-to-end deep neural network models for ECG-based authentication are proposed. In the first model, a convolutional neural network (CNN) is developed and in the second model, a residual convolutional neural network (ResNet) with attention mechanism called ResNet-Attention is designed for human authentication. We have used 2-s duration ECG signals obtained from two ECG databases (Physikalisch-Technische Bundesanstalt [PTB] and Check Your Bio-signals Here initiative [CYBHi]) for authentication. Our proposed ResNet-Attention algorithm achieved an accuracy of 98.85 and 99.27% using PTB and CYBHi, respectively. The results obtained by our developed model show that the performance is better than existing algorithms and can be used in real-time authentication systems after the validation with more diverse ECG data. © 2020 John Wiley & Sons, Ltd

关键词： Electrocardiograms

来源：评论

学校读者我要写书评

暂无评论

Detection of Valvular Heart Diseases Using Fourier Transform and Simple CNN Model

IAENG International Journal of Computer Science

引用

IAENG International Journal of computer Science 2022年第4期49卷 985页

作者： Al-Sharu, Wafaa N. Member, Ali Mohammad Alqudah Qazan, Shoroq Alqudah, Amin Department of Electrical Engineering Hashemite University P.O. Box 330127 Zarqa13133 Jordan Department of Biomedical Systems and Informatics Engineering Yarmouk University P.O Box 566 Irbid21163 Jordan Department of Computer Engineering Yarmouk University P.O Box 566 Irbid21163 Jordan

In recent years, automated diagnosis of the state of health of the heart, particularly cardiac valves, has gained great success using the phonocardiogram (PCG). This work provides a low-complexity, completely automated system for diagnosing and categorizing cardiac illness based on the direct application of a multiclass Convolutional Neural Network (CNN) model, either using Softmax Classifier or KNN or SVM as a classification layer to the fast Fourier transform (FFT) of PCG signals. PCG signals are supplied into the CNN and transformed from the time domain to the frequency domain. With an analysis time of fewer than 2 seconds, the suggested technology allows us to improve performance by up to 97.66%. In the second evaluation, the methodology was evaluated on PhysioNet/Computing in Cardiology Challenge 2016 dataset achieved very high accuracy. © 2022, IAENG International Journal of computer Science. All Rights Reserved.

关键词： Fast Fourier transforms

来源：评论

学校读者我要写书评

暂无评论

Revealing Key Amino Acids on HMGB1 Box A Domain which can be used for Peptide Targeting

引用

Neuroscience Applied 2024年 3卷

作者： O.M. Apaydin O. Sensoy S. Baday Istanbul Technical University Institute of Computer Science Department Informatics Istanbul Türkiye Istanbul Technical University Department of Artificial Intelligence and Data Engineering- Faculty of Computer Informatics and Engineering Istanbul Türkiye

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：