检索结果-内蒙古大学图书馆

arXiv 2021年

作者： Chauhan, Vinod Kumar Singh, Sukhdeep Sharma, Anuj Department of Engineering Science University of Oxford UK United Kingdom D.M. College Panjab University Chandigarh Punjab Moga India Department of Computer Science and Applications Panjab University Chandigarh India

Handwritten character recognition (HCR) remains a challenging pattern recognition problem despite decades of research, and lacks research on script independent recognition techniques. This is mainly because of similar character structures, different handwriting styles, diverse scripts, handcrafted feature extraction techniques, unavailability of data and code, and the development of script-specific deep learning techniques. To address these limitations, we have proposed a script independent deep learning network for HCR research, called HCR-Net, that sets a new research direction for the field. HCR-Net is based on a novel transfer learning approach for HCR, which partly utilizes feature extraction layers of a pre-trained network. Due to transfer learning and image augmentation, HCR-Net provides faster and computationally efficient training, better performance and generalizations, and can work with small datasets. HCR-Net is extensively evaluated on 40 publicly available datasets of Bangla, Punjabi, Hindi, English, Swedish, Urdu, Farsi, Tibetan, Kannada, Malayalam, Telugu, Marathi, Nepali and Arabic languages, and established 26 new benchmark results while performed close to the best results in the rest cases. HCR-Net showed performance improvements up to 11% against the existing results and achieved a fast convergence rate showing up to 99% of final performance in the very first epoch. HCR-Net significantly outperformed the state-of-the-art transfer learning techniques and also reduced the number of trainable parameters by 34% as compared with the corresponding pre-trained network. To facilitate reproducibility and further advancements of HCR research, the complete code is publicly released at https://***/jmdvinodjmd/HCR-Net. Copyright © 2021, The Authors. All rights reserved.

关键词： Character recognition

来源：评论

学校读者我要写书评

暂无评论

Explainable AI for Fraud Detection: Enhancing Transparency and Trust in Financial Decision-Making

Explainable AI for Fraud Detection: Enhancing Transparency a...

引用

Artificial Intelligence in Education and Industry 4.0 (IDICAIEI), DMIHER International Conference on

作者： Rahul Kapale Prasanna Deshpande Shyam Shukla Shailesh Kediya Yoginee Pethe Sujit Metre Priyadarshini Lokmanya Tilak Institute of Management Studies and Research Nagpur India School of Ports Terminal Management and Logistics Symbiosis Skills and Professional University Pune India BVDU IMED Pune India School of Logistics and Supply Chain Management Symbiosis Skills and Professional University Pune India Department of Computer Science and Applications Ramdeobaba University Nagpur India S. B. City College Nagpur India

ISBN: (数字)9798331528713

ISBN: (纸本)9798331528720

Within the past few a long time, utilizing Counterfeit Insights (AI) to discover extortion in managing an account frameworks has gotten a parcel of consideration since it seem make things more secure and more effective. But when AI models are utilized, particularly in critical areas like keeping money, there has to be a adjust between exactness and openness. This exposition looks at how Logical AI (XAI) can offer assistance unravel these issues by making trick discovery frameworks less demanding to understand. This will lead to more openness and trust within the ways that monetary choices are made. The most reason of this think about is to see into how XAI strategies can offer assistance us get it how complex AI models utilized for trick spotting truly work. By clarifying how these models make choices, XAI not as it were increments openness but too makes a difference individuals who have a stake within the matter, like budgetary controllers, bookkeepers, and conclusion clients, get it why certain choices are made. This information is exceptionally critical for making beyond any doubt that individuals are held dependable and take after the rules set by the Basel Committee on Managing an account Supervision (BCBS) and the Common Information Security Control (GDPR). The think about moreover looks at diverse XAI strategies and how they can be utilized to find fraud. These incorporate model-agnostic approaches such as LIME (Nearby Interpretable Model-agnostic Clarifications) and SHAP (SHapley Added substance Clarifications). These strategies grant you data around how critical highlights are by appearing you which components have the greatest affect on foreseeing extortion. These sorts of findings not as it were offer assistance to test and progress models, but they moreover allow money related educate more data to assist them make superior choices approximately how to handle hazard and halt tricks. This consider moreover talks about the benefits of XAI that go beyond jus

关键词： Industries Explainable AI Decision making Information security Medical services Hazards Fraud Reliability Iterative methods Guidelines

来源：评论

学校读者我要写书评

暂无评论

Enhancing fault detection and classification in optical fiber networks with deep learning algorithms

引用

AIP Conference Proceedings 2024年第1期3209卷

作者： Shubham Kaushik Kanwal Garg Jyoti Verma Isha Kansal Preeti Sharma Vikas Khullar Monika Sethi Punjab Engineering College (Deemed-to-be University) Chandigarh India Department of Computer Science and Applications Kurukshetra University Kurukshetra Haryana 136119 India. Department of Computer Science and Engineering Panjabi University Patiala India Chitkara University Institute of Engineering and Technology Chitkara University Punjab India

The detection and classification of faults in optical fiber networks are essential for maintaining their performance and uninterrupted service, as they are vital communication infrastructures. The study attempts to assess the efficacy of deep learning (DL) architectures, employing Dense, Bidirectional-Long Short Term Memory (BiLSTM-Dense), LSTM-Dense, (Convolutional Neural Network) CNN-LSTM, and CNN-BiLSTM, in terms of performance. The findings indicate that Hist Gradient Boosting, Random Forest, and Extra Trees demonstrate exceptional precision in detecting and categorizing faults. CNN-BiLSTM and CNN-LSTM are considered the most effective DL models, as they exhibit exceptional performance by attaining high accuracy, precision, and recall levels. This study offers valuable contributions by shedding light on the appropriateness of various models for detecting and classifying faults in optical fiber networks. As a result, it guides both researchers and practitioners in selecting the most suitable models for their respective applications. The present study addresses the necessity for comparative analysis and evaluation of diverse models. By evaluating diverse machine learning and DL models, scholars and professionals can comprehend the advantages and limitations of each methodology. This data facilitates choosing the most appropriate models for optical fiber networks’ fault detection and classification assignments.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Study of Intrusion Detection System using Machine Learning Classification Algorithm based on different feature selection approach 3

A Study of Intrusion Detection System using Machine Learning...

引用

3rd International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud), I-SMAC 2019

作者： Illavarason, P. Kamachi Sundaram, B. Periyar Maniammai Institute of Science and Technology Department of Computer Science and Applications Thanjavur India

ISBN: (纸本)9781728143651

Network security is the most challenging task of the modern digital era. Due to the development in internet, the number of network attacks has also increased, this is prevented by access control, key manager, and intrusion detection system. Among these the most challenging task is intrusion detection system that ensures the network security. The current approach focuses on the important issues in intrusion detection system, which will identify the unwanted attacks and unauthorized access in the network. The comprehensive overview of the detailed survey is analyzed with the existing data set for identifying the unusual attacks that can understand the current issues in intrusion detection problems. The detailed investigation is reported for observing several issues on the intrusive performance by using the machine learning classification. Here machine learning classification algorithm is used for detecting the several category of attacks. Furthermore, this study evaluates the performance criteria based on the feature extraction and machine learning classification techniques algorithm. Finally, based on the results observed we recommend some important features by using machine learning classification in order to find out the efficient method for detecting the particular attack. © 2019 IEEE.

关键词： Intrusion detection

来源：评论

学校读者我要写书评

暂无评论

High-order thread-safe lattice Boltzmann model for HPC turbulent flow simulations

arXiv

引用

arXiv 2024年

作者： Montessori, Andrea Rocca, Michele La Amati, Giorgio Lauricella, Marco Tiribocchi, Adriano Succi, Sauro Department of Civil Computer Science and Aeronautical Technologies Engineering Rome00185 Italy SCAI Super Computing Applications and Innovation Department CINECA Via dei Tizii 6 Rome00185 Italy Istituto per le Applicazioni del Calcolo CNR via dei Taurini 19 Rome00185 Italy INFN "Tor Vergata" Via della Ricerca Scientifica 1 Roma00133 Italy Center for Life Nano Science@La Sapienza Istituto Italiano di Tecnologia Roma00161 Italy Department of Physics Harvard University CambridgeMA02138 United States

We present a highly-optimized thread-safe lattice Boltzmann model in which the non-equilibrium part of the distribution function is locally reconstructed via recursivity of Hermite polynomials. Such a procedure allows the explicit incorporation of non-equilibrium moments of the distribution up to the order supported by the lattice. Thus, the proposed approach increases accuracy and stability at low viscosities without compromising performances and amenability to parallelization with respect to standard lattice Boltzmann models. The high-order thread-safe version, successfully employed here to simulate the flow in a smooth straight channel at Reτ = 180 and the axisymmetric turbulent jet at Re = 7000, a) achieves peak performances (∼ 5 TeraFlop/s and an arithmetic intensity of ∼ 7 FLOP/byte on single GPU) by significantly reducing the memory footprint, b) retains the algorithmic simplicity of standard lattice Boltzmann computing and c) allows to perform stable simulations at vanishingly low viscosities. Our findings open attractive prospects for high-performance simulations of realistic turbulent flows on GPU-based architectures. Such expectations are confirmed by the excellent agreement among lattice Boltzmann, experimental, and DNS reference data. © 2024, CC BY.

关键词： Turbulent flow

来源：评论

学校读者我要写书评

暂无评论

Preface

Lecture Notes in Electrical Engineering

引用

Lecture Notes in Electrical Engineering 2024年 1258卷 v-ix页

作者： Jain, Sarika Mihindukulasooriya, Nandana Janev, Valentina Shimizu, Cogan Matthew Department of Computer Applications National Institute of Technology Kurukshetra India MIT-IBM Watson AI Lab CambridgeMA United States Institute Mihajlo Pupin Belgrade Serbia Department of Computer Science Wright State University Dayton United States

来源：评论

学校读者我要写书评

暂无评论

The Estimation of PR Method with the use of LV of Web Pages

The Estimation of PR Method with the use of LV of Web Pages

引用

Advance Computing and Innovative Technologies in Engineering (ICACITE), International Conference on

作者： G Tarun Vikrant Sharma Mustafa Mudhafar Vipluv Pathak Qasim Abdul Jabbar Adil Abbas Alwan Laith H. Alzubaidi Ali Saad Alwan Department of computer applications Hindustan institute of technology and science Department of CSE Graphic Era Hill University Adjunct Professor Graphic Era Deemed to be University Dehradun Uttarakhand Department of Anesthesia Techniques and Intensive Care Al-Taff university college Kerbala Iraq Department of Management Studies GL Bajaj Institute of Technology and Management Greater Noida India Medical Instrument Department Alkunooze University College Basra Iraq College of Technical Engineering National University of Science and Technology Dhi Qar Iraq Department of computers Techniques engineering College of technical engineering The Islamic university Najaf Iraq Business Administration Department Kut University College Al kut wasit Iraq

ISBN: (数字)9798350360165

ISBN: (纸本)9798350360172

The fast expansion of the World Wide Web has made it increasingly harder to navigate through its tremendous content, necessitating green and effective internet net page rating algorithms for serps like google. Traditional rating algorithms, which incorporates PageRank, HITS, and Weighted PageRank, drastically talking rent hyperlink-primarily based metrics to decide the relevance and importance of internet pages. However, those algorithms frequently fail to include person engagement statistics, which can provide important insights into the real software program of net pages. This paper introduces a completely unique approach to web page ranking-the Link-Visit PageRank (LVP) set of rules-which enhances traditional models through way of incorporating the frequency of link visits. By integrating how often customers definitely click on on links, the LVP algorithm targets to supply greater correct and customer-centric web page scores. This approach now not only ensures to decorate the relevance of seek engine consequences but additionally aligns with the dynamic and consumer-pushed nature of contemporary web navigation.

关键词： Measurement Navigation Heuristic algorithms Software algorithms Web pages Estimation Software

来源：评论

学校读者我要写书评

暂无评论

Optimize the translation of Indian languages for large target vocabulary in neural machine translation using multi-head attention mechanism

引用

AIP Conference Proceedings 2023年第1期2904卷

作者： Rony Tom Juby Mathew Department of Computer Applications Amal Jyothi College of Engineering Kanjirappally 686518 Kottayam Kerala India Department of Computer Science & Engineering Amal Jyothi College of Engineering Kanjirappally 686518 Kottayam Kerala India

A popular approach in machine translation is neural machine translation. It predicts and translates word sequences using artificial neural networks, generally modelling full sentences in a single integrated model. It enables the translation of languages from one natural language into another with or without the assistance of human beings. Even though deep neural architecture-based machine translation produces accurate results when translating western languages, we are unable to straight away apply these algorithms to translate languages in India, primarily because good corpora are not readily available and our languages are generally morphologically bountiful in nature. This paper aims at providing a general overview on language translation, which consists of Neural Machine Translation and Natural Language Generation involving the translation of Indian languages.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Unextendible and Strongly Uncompletable Product Bases

arXiv

引用

arXiv 2024年

作者： Zhen, Xiao-Fan Zuo, Hui-Juan Shi, Fei Fei, Shao-Ming School of Mathematical Sciences Hebei Normal University Shijiazhuang050024 China State Key Laboratory of Networking and Switching Technology Beijing University of Posts and Telecommunications Beijing100876 China Hebei Key Laboratory of Computational Mathematics and Applications Shijiazhuang050024 China Hebei International Joint Research Center for Mathematics and Interdisciplinary Science Shijiazhuang050024 China Department of Computer Science School of Computing and Data Science University of Hong Kong Hong Kong999077 Hong Kong School of Mathematical Sciences Capital Normal University Beijing10048 China

In 2003, DiVincenzo et al. put forward the question that whether there exists an unextendible product basis (UPB) which is an uncompletable product basis (UCPB) in every bipartition [DiVincenzo et al. Commun. Math. Phys. 238, 379-410(2003)]. Recently, Shi et al. presented a UPB in tripartite systems that is also a strongly uncompletable product basis (SUCPB) in every bipartition [Shi et al. New J. Phys. 24, 113-025 (2022)]. However, whether there exist UPBs that are SUCPBs in only one or two bipartitions remains unknown. We provide a sufficient condition for the existence of SUCPBs based on a quasi U-tile structure. We analyze all possible cases about the relationship between UPBs and SUCPBs in tripartite systems. In particular, we construct a UPB with smaller size d3 − 3d2 + 1 in Cd Cd Cd, which is an SUCPB in every bipartition and has a smaller cardinality than the existing one. Copyright © 2024, The Authors. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Forecasting Solar Energy Generation and Load Consumption - A Method to select the forecasting model based on data type 2

Forecasting Solar Energy Generation and Load Consumption - A...

引用

2nd International Conference on Intelligent Computing, Instrumentation and Control Technologies, ICICICT 2019

作者： Lakshmi Nambiar, M. Krishna Chandran, K.S. Geethalekshmy, V. Mohan, Akshay Amrita Vishwa Vidyapeetham Department of Computer Science and Engineering Amritapuri India Amrita Vishwa Vidyapeetham Department of Computer Science and Applications Amritapuri India

In this study load consumption and solar energy generation data are used to create a model for forecasting. Different techniques like statistical models, machine and deep learning algorithms are used in this study. Th... 详细信息

ISBN: (纸本)9781728102832

关键词： Long short-term memory

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：