检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Mishne, Gal Charles, Adam Halıcıoğlu Data Science Institute Department of Electrical and Computer Engineering the Neurosciences Graduate Program UC San Diego 9500 Gilman Drive La Jolla CA92093 United States Department of Biomedical Engineering Kavli Neuroscience Discovery Institute Center for Imaging Science Department of Neuroscience Mathematical Institute for Data Science Johns Hopkins University BaltimoreMD21287 United States

Optical imaging of the brain has expanded dramatically in the past two decades. New optics, indicators, and experimental paradigms are now enabling in-vivo imaging from the synaptic to the cortex-wide scales. To match the resulting flood of data across scales, computational methods are continuously being developed to meet the need of extracting biologically relevant information. In this pursuit, challenges arise in some domains (e.g., SNR and resolution limits in micron-scale data) that require specialized algorithms. These algorithms can, for example, make use of state-of-the-art machine learning to maximally learn the details of a given scale to optimize the processing pipeline. In contrast, other methods, however, such as graph signal processing, seek to abstract away from some of the details that are scale-specific to provide solutions to specific sub-problems common across scales of neuroimaging. Here we discuss limitations and tradeoffs in algorithmic design with the goal of identifying how data quality and variability can hamper algorithm use and dissemination. © 2024, CC BY.

关键词： Neuroimaging

来源：评论

学校读者我要写书评

暂无评论

Adaptive Methods for Kernel Initialization of Convolutional Neural Network Model Applied to Plant Disease Classification 11

Adaptive Methods for Kernel Initialization of Convolutional ...

引用

11th IEEE Open Conference of Electrical, Electronic and Information sciences, eStream 2024

作者： Lagnaoui, Saloua Boumais, Khaoula Fallah, Saad El En-Naimani, Zakariae Haddouch, Khalid Matuzevicius, Dalius ENSA Sidi Mohamed Ben Abdellah University Laboratory Engineering Systems and Applications Fez Morocco Sidi Mohammed Ben Abdallah University ENSA Laboratory AI Data Science and Emerging Systems Fez Morocco Private University of Fez Systems and Sustainable Environment Laboratory Faculty of Engineering Sciences Fez Morocco ENSET Laboratory Computer Science AI and Cyber Security University of Hassan II Casablanca Mohammedia Morocco Vilnius Gediminas Technical University Faculty of Electronics Department of Electronic Systems Vilnius Lithuania

ISBN: (纸本)9798350352412

Convolutional Neural Networks are instrumental in artificial intelligence, especially in image processing, where their ability to autonomously learn hierarchical features has led to significant breakthroughs. However, the success of these models is intricately tied to the judicious choice of hyperparameters, which include the configuration of convolutional layers, activation functions, and kernel initialization methods. This research explores kernel initialization methods in Convolutional Neural Network models, seven diverse initialization methods (Glorot Uniform, Ones initialization, Zero initialization, Constant initialization, Random initialization, HeNormal initialization, and Orthogonal initialization) are comprehensively compared. The primary objective is to showcase the sensitivity of Convolutional Neural Networks to these various initialization techniques. The study not only aims to reveal the nuanced impact of kernel initialization but also introduces an adaptive method to enhance model performance. By delving into the intricacies of initialization methods, this research contributes to the improvement of Convolutional Neural Networks effectiveness, especially in critical applications like Plant Disease classification. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Shielding NLP Systems: An In-depth Survey on Advanced AI Techniques for Adversarial Attack Detection in Cyber Security

Shielding NLP Systems: An In-depth Survey on Advanced AI Tec...

引用

Automation, Computing and Renewable Systems (ICACRS), International Conference on

作者： T. Sai Lalith Prasad K. B. Manikandan J. Vinoj Research Scholar Department of Computer Science and Engineering Vignan's Foundation for Science and Technology & Research (Deemed to be University) Guntur Andhra Pradesh India Assistant Professor Department of Artificial Intelligence and Data Science Vignan Institute of Technology and Science Telangana India Assistant Professor Department of CSE Vignan's Foundation for Science Technology and Research (Deemed to be University) Guntur Andhra Pradesh India

ISBN: (数字)9798331532420

ISBN: (纸本)9798331532437

This research investigates the challenges posed by adversarial perturbations to the robustness and reliability of Natural Language Processing (NLP) models. The study focuses on three key paradigms for enhancing the resilience of NLP systems: meta-learning, reinforcement learning, and self-supervised learning. A comprehensive evaluation of these strategies is conducted, analyzing their strengths and weaknesses in mitigating adversarial attacks. The study provides insights into the most promising approaches for enhancing the robustness of NLP models and offers directions for future research in this critical area. By advancing the resilience of NLP systems, this research contributes to the development of more robust and trustworthy AI applications across various domains, including healthcare, finance, and security.

关键词： Training Surveys Renewable energy sources Scalability Self-supervised learning Natural language processing Robustness Security Artificial intelligence Resilience

来源：评论

学校读者我要写书评

暂无评论

Understanding the Limitations of Secure Shell (SSH) in Wireless Network Security

Understanding the Limitations of Secure Shell (SSH) in Wirel...

引用

Smart Generation Computing, Communication and Networking (SMART GENCON), International Conference on

作者： Amit Upadhyay Sandip Kulkarni Sharmistha Roy K. Yuvaraj V. Ashok Sunil D. Kale Department of Computer Science Mangalayatan University Aligarh India Department of Computer Science Himalayan University Itanagar India Department of Coputing & Information Technology Usha Martin University Ranchi India Department of Computer Science Karpagam Academy of Higher Education Coimbatore Department of Mechanical Engineering Prince Shri Venkateshwara Padmavathy Engineering College Chennai Department of Artificial Intelligence & Data Science Vishwakarma Institute of Information Technology Pune INDIA

comfy Shell (SSH) is an effective community protocol used to securely connect devices over a computer community. It gives information encryption for relaxed conversation among clients and servers, and can be used to guard information at the same time as being transmitted over a wi-fi community. But, as with every protocol, there are positive obstacles that should be considered when the usage of SSH to relaxed wi-fi networks. Because of the character of wireless networks, in which records is transmitted through the air, SSH is at risk of numerous vulnerabilities. First, SSH is susceptible to man-in-the- center attacks, which can permit an attacker to intercept the network traffic and examine, alter, or block positive packets. Specially, using SSH makes it viable for an attacker to perform password sniffing, for the reason that protocol does not encrypt the user's login credentials. Additionally, the usage of SSH protocols does now not shield against malware or different malicious software, making it feasible for an attacker to monitor or manipulate an affected system.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Deep Learning Based Abstractive Text Summarization: A Survey

Deep Learning Based Abstractive Text Summarization: A Survey

引用

engineering and Technology (PICET), Parul International Conference on

作者： G Uday Kiran Ramakrishna Gandi M. Lavanya Girish Bhagwant Desale Ch. Upendar Rao B. Veerasekhar Reddy Department of CSE(AI & ML) B V Raju Institute of Technology Narsapur India Department of Computer Science & Engineering Madanapalle Institute of Technology & Science India Department of CSE(Data Science) Vignana Bharathi Institute of Technology Hyderabad India Department of Computer Science & IT JET'S Z. B. Patil College DHULE M.S. Department of Information Technology MLR Institute of Technology Hyderabad India

ISBN: (数字)9798350369748

ISBN: (纸本)9798350369755

The explosion in the amount of textual data in the last several years has created an abundant source of resources for data extraction and analysis. It is necessary to summarize this data in order to access relevant information in a fair amount of time. Here we take a look at some of the more current methods for summarizing abstractive texts with deep learning models. A hotspot for research in the field of Natural Language Processing (NLP) is summarization, the process of condensing a document without losing any of its significance. You can classify summarization methods as either “extractive” or “abstractive” depending on whether they use natural language processing to construct new phrases or keep the exact ones from the source text. Extensive research has been conducted on extractive summarization, and the issue has now attained maturity. At this point, the focus of the study is on providing an abstract summary. Abstractive Text Summarization (ATS) is a tough and complicated operation because of the inherent complexity of the natural language content. Researchers can get a thorough overview of Deep Learning based A TS in this work. This paper provides a comprehensive survey of deep learning models employed for abstractive summarization, delving into their architectures and methodologies. Furthermore, it scrutinizes prevalent datasets utilized for training and evaluation purposes, offering insights into their characteristics and suitability for benchmarking. The study evaluates the performance of various abstractive summarization methods on these datasets, shedding light on their effectiveness and limitations. Lastly, emerging themes in the field are discussed, along with unresolved challenges, aiming to foster deeper understanding and advancement in DL-based abstractive text summarization among researchers.

关键词： Deep learning Surveys Measurement Training Text summarization Linguistics Transformers Natural language processing Long short term memory Postal services

来源：评论

学校读者我要写书评

暂无评论

Virtual Palette: An Efficient Object Tracking Tool

Virtual Palette: An Efficient Object Tracking Tool

引用

Sustainable Computing and Smart Systems (ICSCSS), International Conference on

作者： R. Madhura S. Nikkath Bushra Devi P. P Sasigresa P Department of Computing Technologies SRM Institute of Science and Technology Kattangulathur Chennai Tamil Nadu India Department of Artificial Intelligence and Machine Learning Vel Tech Rangarajan Dr. Sagunthala R&D Institute of Science and Technology Department of Computer Science and Engineering (Data Science) Vel Tech Rangarajan Dr. Sagunthala R&D Institute of Science and Technology

ISBN: (数字)9798350379990

ISBN: (纸本)9798350391558

Virtual Palette is a cutting-edge tool designed to enhance audience participation by providing an alternative to conventional jamboards for educators. Leveraging object tracking, a fundamental component of computer vision, it facilitates seamless and intuitive interactions with the audience. The video analysis process comprises three primary steps: object detection, object tracking across frames, and object behavior analysis. This sophisticated approach necessitates meticulous consideration of elements such as accurate object representation, feature selection for tracking, object recognition, and object tracking. One of the primary advantages of Virtual Palette over traditional mouse or touchpad devices is its ability to support clear and legible drawing and effective interaction in virtual environments. This feature is particularly beneficial for individuals with hearing impairments, aiding in the identification of patterns and drawings. The development of Virtual Palette involves overcoming challenges such as precise three-dimensional finger movement tracking and optimizing the technology for virtual backgrounds. Overall, Virtual Palette offers a dynamic and interactive teaching experience, akin to the advantages provided by screen recording for creating tutorials and virtual documentation.

关键词： Fingers Virtual environments Tutorials Transforms Object detection Green computing Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Deep Learning Approach for Measuring the Severity of Depression Using Social Media

Deep Learning Approach for Measuring the Severity of Depress...

引用

2023 International Conference on Integration of Computational Intelligent System, ICICIS 2023

作者： Sable, Nilesh P. Patil, Rajkumar Gaikwad, Vidya Deshmukh, Jyoti Yogesh Buchade, Amar Pune India Design & Technology University Deaprtment of Infotmation Technology Mit Art Pune India Bansilal Ramnath Agarwal Charitable Trust's Vishwakarma Institute of Information Technology Department of Computer Engineering Pune India Marathwada Mitramandal's Institute of Technology Department of Computer Engineering Pune India Bansilal Ramnath Agarwal Charitable Trust's Vishwakarma Institute of Information Technology Department of Artificial Intelligence & Data Science Pune India

ISBN: (纸本)9798350318784

Depression., one of the world's most common medical illnesses today, has a negative social and financial impact. The emotional computing models concentrate more attention on analyzing specific client postings than chronological conduct patterns, which are essential for spotting changes and improvements in psychological behavior and user behavior. Context, which entails a consistent examination of emotions and traits over time. This article suggests using the TROAD framework to recognize sequential trends over time in depressive user behavior on social media. To recognize sequential patterns from depressive user behavior, the framework determines the optimal interval to analyze each user activity, collects emotional and contextual variables from user input, and models the features into time frames. Today, depression has become a significant issue in our culture. It is a significant contributor to suicide, particularly among teenagers. The nations impacted by the current coronavirus disease outbreak (COVID-19) have suggested lockdown and social seclusion measures. These policies, which lead to interpersonal isolation, have caused depression and mental health issues to become very major issues. Clinical psychologists typically conduct in-person interviews with sad individuals to provide a diagnosis by the clinical depression criteria. However, individuals frequently avoid seeing doctors when they are first experiencing symptoms of depression. Nowadays, more and more people communicate their sentiments on social media. © 2023 IEEE.

关键词： Coronavirus

来源：评论

学校读者我要写书评

暂无评论

A Cost-Effective Smart Helmet for Human Safety and Road Accident Detection Using loT

A Cost-Effective Smart Helmet for Human Safety and Road Acci...

引用

Recent Progresses in science, engineering and Technology (ICRPSET), International Conference on

作者： Sharmin Akter Md Abu Yousuf K. M. Rafiqul Alam Md. Sahidullah Md. Monirul Islam Jia Uddin Computer Science and Engineering Atish Dipankar University of Science and Technology Dhaka Bangladesh Department of Electrical Engineering Wright State University Dayton Ohio USA Department of Statistics Jahangirnagar University Dhaka bangladesh Computer Science and Engineering Asian University of Bangladesh (AUB) Dhaka Bangladesh Information and Communications Engineering Hankuk University of Foreign Studies Gyeonggi-do South Korea AI and Big Data Department Woosong University Daejeon South Korea

ISBN: (数字)9798331509477

ISBN: (纸本)9798331509484

A smart helmet is a type of protective headgear used by the rider, which makes bike driving safer than before. Nowadays road accidents occur very frequently in our daily lives because motorcyclists are not always wearing helmets. To avoid road accidents, motorcyclists have to wear helmets. By wearing it, riders can save their life up to some instant. In this paper, we proposed a smart helmet system based on Arduino and loT devices named Ultrasonic sensor, Vibration sensor, GPS module, MQ-3 Alcohol sensor, Wi-Fi module, IR sensor, and Relay module. This system aims to improve the safety of riders. The design of this smart helmet is to detect whether the user wears a helmet or not with the help of an IR sensor. Apart from this, this smart system can detect whether the user is an alcoholic or not. If the user is alcoholic, the ignition of the bike gets OFF. Only the ignition gets ON while the users wear a helmet and do not drink alcohol. It will benefit riders and medical practitioners when an accident occurs with immediate treatment. This will bring more safety to bike riding and will save many lives as well.

关键词： Vibrations Head Road accidents Microcontrollers Motorcycles Ignition Smart systems Safety Relays Wireless fidelity

来源：评论

学校读者我要写书评

暂无评论

Bioinformatic Analysis Reveals both Oversampled and Underexplored Biosynthetic Diversity in Nonribosomal Peptides

引用

ACS Chemical Biology 2023年第3期18卷 476-483页

作者： Jian, Bo-Siyuan Chiou, Shao-Lun Hsu, Chun-Chia Ho, Josh Wu, Yu-Wei Chu, John Department of Computer Science and Information Engineering National Taiwan University Taipei 10617 Taiwan Department of Chemistry National Taiwan University Taipei 10617 Taiwan Graduate Institute of Biomedical Informatics College of Medical Science and Technology Taipei Medical University Taipei 10675 Taiwan Clinical Big Data Research Center Taipei Medical University Hospital Taipei 10675 Taiwan TMU Research Center for Digestive Medicine Taipei Medical University Taipei 10675 Taiwan

The traditional natural product discovery approach has accessed only a fraction of the chemical diversity in nature. The use of bioinformatic tools to interpret the instructions encoded in microbial biosynthetic genes has the potential to circumvent the existing methodological bottlenecks and greatly expand the scope of discovery. Structural prediction algorithms for nonribosomal peptides (NRPs), the largest family of microbial natural products, lie at the heart of this new approach. To understand the scope and limitation of the existing prediction algorithms, we evaluated their performances on NRP synthetase biosynthetic gene clusters. Our systematic analysis shows that the NRP biosynthetic landscape is uneven. Phenylglycine and its derivatives as a group of NRP building blocks (BBs), for example, have been oversampled, reflecting an extensive historical interest in the glycopeptide antibiotics family. In contrast, the benzoyl BB, including 2,3-dihydroxybenzoate (DHB), has been the most underexplored, hinting at the possibility of a reservoir of as yet unknown DHB containing NRPs with functional roles other than a siderophore. Our results also suggest that there is still vast unexplored biosynthetic diversity in nature, and the analysis presented herein shall help guide and strategize future natural product discovery campaigns. We also discuss possible ways bioinformaticians and biochemists could work together to improve the existing prediction algorithms. © 2023 The Authors. Published by American Chemical Society.

关键词： Algorithms Bacteria Bioinformatics and computational biology Peptides and proteins Pharmaceuticals

来源：评论

学校读者我要写书评

暂无评论

NN-Steiner: A Mixed Neural-algorithmic Approach for the Rectilinear Steiner Minimum Tree Problem

arXiv

引用

arXiv 2023年

作者： Kahng, Andrew B. Nerem, Robert R. Wang, Yusu Yang, Chien-Yi Department of Electrical and Computer Engineering University of California San Diego United States Department of Computer Science and Engineering University of California San Diego United States Halıcıoglu Data Science Institute University of California San Diego United States

Recent years have witnessed rapid advances in the use of neural networks to solve combinatorial optimization problems. Nevertheless, designing the "right" neural model that can effectively handle a given optimization problem can be challenging, and often there is no theoretical understanding or justification of the resulting neural model. In this paper, we focus on the rectilinear Steiner minimum tree (RSMT) problem, which is of critical importance in IC layout design and as a result has attracted numerous heuristic approaches in the VLSI literature. Our contributions are two-fold. On the methodology front, we propose NN-Steiner, which is a novel mixed neural-algorithmic framework for computing RSMTs that leverages the celebrated PTAS algorithmic framework of Arora to solve this problem (and other geometric optimization problems). Our NN-Steiner replaces key algorithmic components within Arora’s PTAS by suitable neural components. In particular, NN-Steiner only needs four neural network (NN) components that are called repeatedly within an algorithmic framework. Crucially, each of the four NN components is only of bounded size independent of input size, and thus easy to train. Furthermore, as the NN component is learning a generic algorithmic step, once learned, the resulting mixed neural-algorithmic framework generalizes to much larger instances not seen in training. Our NN-Steiner, to our best knowledge, is the first neural architecture of bounded size that has capacity to approximately solve RSMT (and variants). On the empirical front, we show how NN-Steiner can be implemented and demonstrate the effectiveness of our resulting approach, especially in terms of generalization, by comparing with state-of-the-art methods (both neural and non-neural based). © 2023, CC BY.

关键词： Combinatorial optimization

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：