检索结果-内蒙古大学图书馆

22nd IEEE International Symposium on Biomedical Imaging, ISBI 2025

作者： Saha, Rekha Saha, Debbrata K. Fu, Zening Adali, Tulay Calhoun, Vince D. Georgia State University Georgia Institute of Technology Emory University Tri-institutional Center for Translational Research in Neuroimaging and Data Science (TReNDS) Atlanta 30303 GA United States University of Maryland Department of Computer Science and Electrical Engineering Baltimore County 21250 MD United States

ISBN: (纸本)9798331520526

Functional network connectivity (FNC) is a valuable measure for assessing the temporal interdependence and intrinsic functional relationships among brain networks. While longitudinal research on intrinsic functional connectivity has been useful for characterizing age-related changes across whole-brain networks, there has been limited focus on capturing multivariate patterns of FNC changes during brain development. In this study, we employed a previously proposed method that utilizes FNC matrices to estimate multiple overlapping brain functional change patterns (FCPs). This method was applied to the extensive Adolescent Brain and Cognitive Development (ABCD) dataset. Our findings reveal several well-structured FCPs indicating significant changes over a four-year period, including brain functional connectivity between visual network (VSN) and sensorimotor network (SMN) domains. This FNC expression pattern strengthens with age. Additionally, we observed a distinct pattern of changes between male and female individuals for two components including males show stronger increasing coupling between the default mode network (DMN) and VSN compared to females. © 2025 IEEE.

关键词： Delta FNC FCP fMRI ICA Longitudinal study

来源：评论

学校读者我要写书评

暂无评论

A Survey of Vietnamese Automatic Speech Recognition 9

A Survey of Vietnamese Automatic Speech Recognition

引用

9th International Conference on Orange Technology, ICOT 2021

作者： Nga, Cao Hong Li, Chung-Ting Li, Yung-Hui Wang, Jia-Ching National Central University Department of Computer Science and Information Engineering Taiwan Hon Hai Research Institute Ai Research Center Taiwan

ISBN: (纸本)9781665478427

In this paper, we survey Vietnamese automatic speech recognition (ASR). The objective of this survey is to provide an overview of the current status and remaining challenges of implementing a Vietnamese ASR. Recently, there are some studies on ASR for Vietnamese language;however, these studies encounter some obstacles and the results obtained compared to other languages such as English or Mandarin are lower. With regards to Vietnamese speech recognition, we will examine the methods of building a system along with speech data and text data collection techniques and available speech resources. We review both the methods applied to acoustic modeling and those used to language modeling. In addition, we convey some directions for future research. © 2021 IEEE.

关键词： Speech recognition

来源：评论

学校读者我要写书评

暂无评论

Reveal training performance mystery between Tensor Flow and PyTorch in the single GPU environment

引用

science China(Information sciences) 2022年第1期65卷 147-163页

作者： Hulin Dai Xuan PENG Xuanhua SHI Ligang HE Qian XIONG Hai JIN National Engineering Research Center for Big Data Technology and System Service Computing Technology and System LabSchool of Computer Science and Technology Huazhong University of Science and Technology Department of Computer Science University of Warwick

Deep learning has gained tremendous success in various fields while training deep neural networks(DNNs) is very compute-intensive, which results in numerous deep learning frameworks that aim to offer better usability and higher performance to deep learning practitioners. Tensor Flow and Py Torch are the two most popular frameworks. Tensor Flow is more promising within the industry context, while Py Torch is more appealing in academia. However, these two frameworks differ much owing to the opposite design philosophy:static vs dynamic computation graph. Tensor Flow is regarded as being more performance-friendly as it has more opportunities to perform optimizations with the full view of the computation graph. However, there are also claims that Py Torch is faster than Tensor Flow sometimes, which confuses the end-users on the choice between them. In this paper, we carry out the analytical and experimental analysis to unravel the mystery of comparison in training speed on single-GPU between Tensor Flow and Py Torch. To ensure that our investigation is as comprehensive as possible, we carefully select seven popular neural networks, which cover computer vision, speech recognition, and natural language processing(NLP). The contributions of this work are two-fold. First, we conduct the detailed benchmarking experiments on Tensor Flow and Py Torch and analyze the reasons for their performance difference. This work provides the guidance for the end-users to choose between these two frameworks. Second, we identify some key factors that affect the performance,which can direct the end-users to write their models more efficiently.

关键词： deep learning performance comparison TensorFlow PyTorch

来源：评论

学校读者我要写书评

暂无评论

Analysis of N-Way K-Shot Malware Detection Using Few-Shot Learning 1

引用

International Conference on Cyber Security, Privacy and Networking, ICSPN 2022

作者： Chui, Kwok Tai Gupta, Brij B. Lee, Lap-Kei Torres-Ruiz, Miguel Kowloon Hong Kong Department of Computer Science and Information Engineering International Center for AI and Cyber Security Research and Innovations Asia University Taichung41354 Taiwan Lebanese American University Beirut1102 Lebanon Department of Electronic Engineering and Computer Science School of Science and Technology Hong Kong Metropolitan University Kowloon Hong Kong Instituto Politécnico Nacional Centro de Investigacion en Computacion UPALM-Zacatenco Mexico City07320 Mexico

ISBN: (数字)9783031220180

ISBN: (纸本)9783031220173

Solving machine learning problems with small-scale training datasets becomes an emergent research area to fill the opposite end of big data applications. Attention is drawn to malware detection using few-shot learning, which is typically formulated as N-way K-shot problems. The aims are to reduce the effort in data collection, learn the rare cases, reduce the model complexity, and increase the accuracy of the detection model. The performance of the malware detection model is analyzed with the variation in the number of ways and the number of shots. This facilitates the understanding on the design and formulation of three algorithms namely relation network, prototypical network, and relation network for N-way K-shot problems. Two benchmark datasets are selected for the performance evaluation. Results reveal the general characteristics of the performance of malware detection model with fixed ways and varying shots, and varying ways and fixed shots based on the trends of results of 30 scenarios. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Benchmarking

来源：评论

学校读者我要写书评

暂无评论

Smart Contract Vulnerability Detection Model Based on Siamese Network 7th

Smart Contract Vulnerability Detection Model Based on Siames...

引用

7th International Conference on Smart Computing and Communication, SmartCom 2022

作者： Chen, Weijie Guo, Ran Wang, Guopeng Zhang, Lejun Qiu, Jing Su, Shen Liu, Yuan Xu, Guangxia Chen, Huiling College of Information Engineering Yangzhou University Yangzhou225127 China Cyberspace Institute Advanced Technology Guangzhou University Guangzhou510006 China Engineering Research Center of Integration and Application of Digital Learning Technology Ministry of Education Beijing100039 China Department of Computer Science and Artificial Intelligence Wenzhou University Wenzhou325035 China

ISBN: (纸本)9783031281235

Blockchain is experiencing the transition from the first generation to the second generation, and smart contract is the symbol of the second generation blockchain. Under the background of the explosive growth of the second-generation blockchain platform and applications represented by smart contracts, frequent smart contract vulnerability events seriously threaten the ecological security of the blockchain, reflecting the importance and urgency of smart contract vulnerability detection. In this paper, we proposed a smart contract vulnerability detection method based on a Siamese network. We combined the Siamese network with Long Short-Term Memory (LSTM) Network neural network to complete the task of smart contract vulnerability detection. The Siamese network used in this paper consists of two subnetworks that share the same parameters onto a low dimension and easily separable feature space. Siamese network is now widely used in the field of image similarity and target tracking. In this paper, we improve the Siamese network so that it can be used for smart contract vulnerability detection. By comparing with previous research results, the model has better vulnerability detection performance and a lower false-positive rate. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Smart contract

来源：评论

学校读者我要写书评

暂无评论

Identification of soybean in Argentina using Sentinel-2 composite images

引用

International Journal of Agricultural and Biological engineering 2024年第5期17卷 266-274页

作者： Linsheng Huang Yue Chen Yuhao Pan Zihang Lou Shijun Zheng Xiaoyang Zhang Le Yu Shengwei Liu Dailiang Peng National Engineering Research Center for Agro-Ecological Big Data Analysis&Application Anhui UniversityHefei 230601China Key Laboratory of Digital Earth Science Aerospace Information Research InstituteChinese Academy of SciencesBeijing 100094China Research Area of Ecology and Biodiversity School of Biological SciencesThe University of Hong KongPokfulamHong KongChina College of Environmental and Resource Sciences Zhejiang UniversityHangzhou 310058China Department of Earth System Science Ministry of Education Key Laboratory for Earth System ModelingInstitute for Global Change StudiesTsinghua UniversityBeijing 100084China Geospatial Sciences Center of Excellence Department of Geography&Geospatial SciencesSouth Dakota State UniversityBrookingsSD 57007USA Ministry of Education Ecological Field Station for East Asian Migratory Birds Beijing 100084China Tsinghua University(Department of Earth System Science)-Xi’an Institute of Surveying and Mapping Joint Research Center for Next-Generation Smart Mapping Beijing 100084China School of Surveying and Land Information Engineering Henan Polytechnic UniversityJiaozuo 454000HenanChina

Soybean is one of the most important oil crops, and Argentina is the third-largest soybean producer in the world, accounting for 17% of the global soybean yield. Timely and accurate information on soybean spatial distribution is critical for ensuring global food security. Sentinel-2 multispectral data and machine learning classification models are used to investigate the potential of soybean identification in the early stage of the growing season in Argentina, with the help of Google Earth Engine (GEE). The earliest time window and optimal feature set for soybean identification are explored. Results are as follows: 1) the random forest (RF) classification model demonstrated the highest level of classification accuracy compared to the backpropagation neural network (BPNN), support vector machine (SVM), and naive Bayes (NB) models;2) Soybean can be accurately identified as early as the end of February (filling stage), which is approximately one month before harvest;3) The optimal feature-subset can reduce the amount of input data by 80% while maintaining high classification accuracy. The overall accuracy (OA) of the RF classification model is 85.87%, and the relative error between the estimated soybean planting area and the agricultural statistics is 3.45%. This study provided a high-precision method for early-season identification of soybeans over large scales. The results can provide a data support for early futures trading and agricultural insurance, as well as a reference for policy-making to ensure global soybean food security.

关键词： soybean machine learning time window feature selection Sentinel-2 Google Earth Engine

来源：评论

学校读者我要写书评

暂无评论

An Efficient Detection Model for Smart Contract Reentrancy Vulnerabilities 7th

An Efficient Detection Model for Smart Contract Reentrancy V...

引用

7th International Conference on Smart Computing and Communication, SmartCom 2022

作者： Li, Yuan Guo, Ran Wang, Guopeng Zhang, Lejun Qiu, Jing Su, Shen Liu, Yuan Xu, Guangxia Chen, Huiling College of Information Engineering Yangzhou University Yangzhou225127 China Cyberspace Institute Advanced Technology Guangzhou University Guangzhou510006 China Engineering Research Center of Integration and Application of Digital Learning Technology Ministry of Education Beijing100039 China Department of Computer Science and Artificial Intelligence Wenzhou University Wenzhou325035 China

ISBN: (纸本)9783031281235

We propose a novel smart contract re-entry vulnerability detection model based on BiGAS. The model combines a BiGRU neural network that introduces an attention mechanism with an SVM. We start from the data features of smart contracts, learn the model layer by layer to achieve feature extraction and vulnerability identification, introduce batch normalization, Dropout processing and use improved model classifiers to improve the vulnerability identification accuracy, model convergence speed and generalization capability of smart contracts. We had conducted numerous experiments, and the experimental results showed that BiGAS Detection Model has a strong vulnerability detection ability. The accuracy of vulnerability detection reached 93.24%, and the F1-score was 93.17%. We compared our approach with advanced automated audit tools and other deep learning-based vulnerability detection methods. The conclusion was that our method is significantly better than the existing advanced methods in detecting smart contract reentrancy vulnerabilities. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Smart contract

来源：评论

学校读者我要写书评

暂无评论

Exploring the Influence of Dimensionality Reduction on Anomaly Detection Performance in Multivariate Time Series

arXiv

引用

arXiv 2024年

作者： Altin, Mahsun Cakir, Altan Department of Computer Engineering Istanbul Technical University Sariyer Istanbul34467 Turkey Department of Physics Engineering Istanbul Technical University Sariyer Istanbul34467 Turkey Artificial Intelligence Data Science Research and Application Center Istanbul Technical University Sariyer Istanbul34467 Turkey Parton Big Data Analytics and Consulting ITU Ariteknokent Sariyer Istanbul34467 Turkey Adin.Ai WilmingtonDE19806 United States

This paper presents an extensive empirical study on the integration of dimensionality reduction techniques with advanced unsupervised time series anomaly detection models, focusing on the MUTANT and Anomaly-Transformer models. The study involves a comprehensive evaluation across three different datasets: MSL, SMAP, and SWaT. Each dataset poses unique challenges, allowing for a robust assessment of the models’ capabilities in varied contexts. The dimensionality reduction techniques examined include PCA, UMAP, Random Projection, and t-SNE, each offering distinct advantages in simplifying high-dimensional data. Our findings reveal that dimensionality reduction not only aids in reducing computational complexity but also significantly enhances anomaly detection performance in certain scenarios. Moreover, a remarkable reduction in training times was observed, with reductions by approximately 300% and 650% when dimensionality was halved and minimized to the lowest dimensions, respectively. This efficiency gain underscores the dual benefit of dimensionality reduction in both performance enhancement and operational efficiency. The MUTANT model exhibits notable adaptability, especially with UMAP reduction, while the Anomaly-Transformer demonstrates versatility across various reduction techniques. These insights provide a deeper understanding of the synergistic effects of dimensionality reduction and anomaly detection, contributing valuable perspectives to the field of time series analysis. The study underscores the importance of selecting appropriate dimensionality reduction strategies based on specific model requirements and dataset characteristics, paving the way for more efficient, accurate, and scalable solutions in anomaly detection. © 2024, CC BY.

关键词： Anomaly detection

来源：评论

学校读者我要写书评

暂无评论

Fault Diagnosis and Region Detection of CCD Image in Double Delay Three-Step Autocorrelator

Fault Diagnosis and Region Detection of CCD Image in Double ...

引用

Pattern Recognition and Machine Learning (PRML), IEEE International Conference on

作者： Suicheng Li Zizhou He Wenwen Shen Yanwen Xia Zhigao Ni Xiaowei Zhou Department of Computer Engineering and Science Institute of Computer Application CAEP Sichuan Province China Department of Optical instrument Research Center of Laser Fusion CAEP Sichuan Province China

With the development of national optical experimental facilities, countries worldwide have invested significant amounts of funding into constructing various facilities. Among them, the double delay three-step autocorrelator is an important instrument for measuring the ultrashort pulse laser waveform and is frequently used in laser experiments. Because electronic instruments are too slow to measure directly, we can only use CCD images to reconstruct the laser waveform. Therefore, CCD, in particular, plays a vital role in the closed-loop control system of double delay three-step autocorrelator, and CCD image quality and processing can directly affect the measurement results of the pulse laser. In this paper, we propose an image feature-based CCD fault diagnosis and effective region detection method for the double delay three-step autocorrelator. This fault diagnosis method first calculates the pixel sum of each row and analyzes the primary pixel distribution of each frame which could represents the main features of CCD image. Next, we design a fault diagnosis model to classify those features and pick out the suitable image. Finally, we analyze and process the above image and design a boundary detection algorithm to detect the effective region for further waveform recovery algorithm. In addition, to verify performance of the proposed method, we conduct a significant number of experiments. Through qualitative and quantitative analyses of the experimental results, we verify the effectiveness and robustness of the proposed method for CCD fault diagnosis.

关键词：

来源：评论

学校读者我要写书评

暂无评论

SBA: A Swift and Stealthy Backdoor Attack Framework for Federated Learning

SBA: A Swift and Stealthy Backdoor Attack Framework for Fede...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Junhan Wang Zhangming Wu Zhuoyue Wang Lu Dong School of Cyber Science and Engineering Southeast University Nanjing China Engineering Research Center of Blockchain Application Supervision and Management Southeast University Ministry of Education Department of Electrical Engineering and Computer Sciences University of California Berkeley USA

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Federated Learning (FL) enables collaborative model training while protecting the privacy of individual participants. However, this decentralized framework is vulnerable to certain risks, particularly backdoor attacks. In such attacks, malicious actors embed triggers into the global model, causing incorrect classifications for inputs with specific features. Traditional backdoor attacks typically use uniform triggers, which makes them more likely to be detected. Furthermore, These methods tend to overlook the fact that poisoned samples may cause abnormal confidence levels in non-target classes. These methods also require an extended attack window to achieve the desired success rate, which reduces their efficiency and increases the chances of detection. This paper introduces a novel backdoor attack method called Swift Backdoor Attack (SBA), designed to overcome these challenges. SBA improves both the stealth and success rate of backdoor attacks by generating adaptive triggers tailored to specific inputs. It also incorporates Non-Target Class Suppression (NTCS) loss and an entropy gain mechanism to enhance the attack’s effectiveness. Our approach suppresses unnecessary activations in the neural network when encountering poisoned samples, making the attack harder to be detected. Experimental results demonstrate that this method significantly increases the success rate of backdoor attacks within a shorter attack window, while evading existing detection mechanisms. This provides a more efficient and covert strategy for executing backdoor attacks in FL. Our code is available at https://***/bigcheeses/SBA.

关键词： Training Privacy Codes Federated learning Neural networks Collaboration Signal processing Entropy Robustness Speech processing

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：