检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Zhang, Lei Long, Tao Xu, Weizhen Cui, Laizhong Liu, Jiangchuan The College of Computer Science and Software Engineering Shenzhen University Guangdong Shenzhen518060 China The School of Computing Science Simon Fraser University BurnabyBCV5A 1S6 Canada

viewport prediction is the crucial task for adaptive 360-degree video streaming, as the bitrate control algorithms usually require the knowledge of the user’s viewing portions of the frames. various methods are studied and adopted for viewport prediction from less accurate statistic tools to highly calibrated deep neural networks. Conventionally, it is difficult to implement sophisticated deep learning methods on mobile devices, which have limited computation capability. In this work, we propose an advanced learning-based viewport prediction approach and carefully design it to introduce minimal transmission and computation overhead for mobile terminals. We also propose a model-agnostic meta-learning (MAML) based saliency prediction network trainer, which provides a few-sample fast training solution to obtain the prediction model by utilizing the information from the past models. We further discuss how to integrate this mobile-friendly viewport prediction (MFvP) approach into a typical 360-degree video live streaming system by formulating and solving the bitrate adaptation problem. Extensive experiment results show that our prediction approach can work in real-time for live video streaming and can achieve higher accuracies compared to other existing prediction methods on mobile end, which, together with our bitrate adaptation algorithm, significantly improves the streaming QoE from various *** observe the accuracy of MFvP is 8.1% to 28.7% higher than other algorithms and achieves 3.73% to 14.96% higher average quality level and 49.6% to 74.97% less quality level change than other algorithms. Copyright © 2024, The Authors. All rights reserved.

关键词： Forecasting

来源：评论

学校读者我要写书评

暂无评论

Development of a Machine Learning Model for pulmonary carcinoma Prediction using computer Tomography (CT) Images

Development of a Machine Learning Model for pulmonary carcin...

引用

International Conference on Image Information Processing

作者： Jampani Chandra sekhar Priyanshu singh T. Rajesh Kumar G. Mahendran Katta subba Rao Ramya Maranan Department of Computer Science and Engineering NRI Institute of Technology Guntur Andhra Pradesh India Department of Computer Science and Engineering R.V. College of Engineering Bengaluru Karnataka India Department of Artificial Intelligence and Machine Learning Sri Krishna College of Technology Coimbatore Tamil Nadu India Department of Mechanical Engineering R.M.K Engineering College Kavaraipettai Chennai Tamil Nadu India Department of Computer Science and Engineering B V Raju Institute of Technology Narsapur Medak (District) Telangana India Department of Research and Innovation Saveetha School of Engineering SIMATS Chennai Tamil Nadu India

ISBN: (数字)9798350371406

ISBN: (纸本)9798350371413

In this study, computer tomography (CT) scans are used to build a machine learning model for predicting lung cancer. A dataset of 10,000 CT scan images from 900 patients was used to train a group of machine learning algorithms. The objective is to identify the most reliable algorithm for lung cancer detection. The algorithms CNN, CNN with gradient descent, vGG-19, and vGG-16 have all been used in real-world applications. Each algorithm is trained using the dataset to create a prediction model. Performance metrics including accuracy, precision, specificity, sensitivity, and F-score are taken into account when assessing how effective the algorithms are. The results demonstrate that the CNN model outperforms the competition with a 97% accuracy rate. The superior accuracy, specificity, sensitivity, and F-score of the CNN model enable accurate lung cancer detection in CT scan images. The outcomes demonstrate that, utilising the well-established machine learning paradigm, the CNN algorithm, in particular, has substantial potential for early lung cancer diagnosis and prediction. Correctly identifying lung cancer from CT scan images allows for the implementation of focused treatment programmes and prompt interventions, ultimately leading to better patient outcomes. The study emphasises the requirement for applying machine learning methods to identify relevant characteristics and patterns in CT scan images. Machine learning models can now be used in clinical settings to support the precise detection and treatment of lung cancer. The use of powerful algorithms and extensive datasets allowed for this accomplishment. In order to validate and enhance the models that have been developed, more study is needed in this field. The models will therefore be more useful and suitable for use in actual clinical settings. Utilising larger datasets, deep learning architectures, and other possible machine learning advancements can significantly increase the precision and generalizability of

关键词： Machine learning algorithms sensitivity Computed tomography Computational modeling Lung cancer Machine learning Predictive models

来源：评论

学校读者我要写书评

暂无评论

Implementation of Full Adder Cells for Ultra Low Power Energy Efficient Computing Applications

Implementation of Full Adder Cells for Ultra Low Power Energ...

引用

sustainable Computing and smart systems (ICsCss), International Conference on

作者： Mohammed Mahaboob basha vadde seetharama Rao Lachi Reddy Poreddy v. Madhurima srinivasulu Gundala Ovidiu Petru stan Department of ECE Sreenidhi Institute of Science and Technology India Department of ECE Lakireddy Bali Reddy College of Engineering India Department of ECE S V College of Engineering (Autonomous) Tirupati Andhra Pradesh India Faculty of Automation and Computer Science Technical University of Cluj Napoca Cluj-Napoca Romania

ISBN: (数字)9798350379990

ISBN: (纸本)9798350391558

Energy Efficiency is of critical significance when designing integrated circuits. In the world of advanced integrated circuits that operate in near-threshold voltage is gaining much awareness because of its potential for high-performance and energy-efficient designs. Today's vLsI design places a great deal of importance on static or leaky power usage as technology advances into the nanoscale realm. In this brief, a one-bit full adder cell with reduced transistor count is suggested for lower power application by employing CMOs, Pass Transistor Logic and Transmission Gate schemes with XOR, XNOR and multiplexer logics at different supply voltages for achieving full swing sum and carry output. Post-layout simulation results reveal that, the proposed one-bit full adder design has achieved remarkable improvements compared to other published designs, boasting over 23 % energy savings and over 32% reductions in Energy Delay Product by employing just 10 transistors. The outcomes demonstrate how efficient the full adder cells can serve as an arithmetic circuit for conventional signal processing computing applications in integrated circuits.

关键词： Power demand simulation Logic gates very large scale integration signal processing Threshold voltage silicon

来源：评论

学校读者我要写书评

暂无评论

Two-stage Deep Learning - YouTube video Recommendation Process

Two-Stage Deep Learning - YouTube Video Recommendation Proce...

引用

New Frontiers in Communication, Automation, Management and security (ICCAMs), International Conference on

作者： Cheepu balakrishna D. sharada Mani A basi Reddy P. Chandrakanth R.senthamil selvan Department of Computer Science and Engineering Saispurthi Institute Of Technology Sathupally Mandal India Department of Computer Science and Engineering QIS college of Engineering and Technology Ongole India Department of Computer Science and Engineering School of Computing Mohan Babu University Tirupati India Department of Computer Science and Engineering N.B.K.R.I.S.T Vidyanagar Kota India Department of Electronics and Communication Engineering Annamacharya Institute of Technology and Sciences Tirupati India

ISBN: (数字)9798350317060

ISBN: (纸本)9798350317077

YouTube is one of the most widely used and technically advanced commercial systems of recommendations. It provides an overview of the system and highlights the remarkable gains in efficiency made possible by deep learning. This research uses the traditional “two-stage” framework for separating and retrieving information. First, a profound model for generating candidates ought to be described, and then a model for ranking candidates ought to be outlined, and both of these should be done in that order. The order in which these two tasks ought to be completed is important. In addition, it offers instruction that can be put into practice as well as valuable insights that have been gleaned from the process of developing, iterating, and managing a large-scale recommendation system that has a significant impact on the experience of the user.

关键词： Deep learning Training video on demand Artificial neural networks Watches Web sites security

来源：评论

学校读者我要写书评

暂无评论

Design and development of machine learning based resume ranking system

Global Transitions Proceedings

引用

Global Transitions Proceedings 2022年第2期3卷 371-375页

作者： Tejaswini K Umadevi v shashank M Kadiwal sanjay Revanna Department of Computer Science and Engineering Dr. Ambedkar Institute of Technology Bengaluru 560 056 Karnataka India Department of Computer Science and Engineering B.M.S. College of Engineering Bengaluru 560 019 Karnataka India

Finding acceptable applicants for a vacant job might be a difficult process, especially when there are many prospects. The manual process of screening resumes could stymie the team's efforts to locate the right individual at the right moment. The laborious screening may be greatly aided by an automated technique for screening and ranking applicants. In our work, the top applicants might be rated using content-based suggestion, which uses cosine similarity to find the curriculum vitae that are the most comparable to the job description supplied and KNN algorithm is used to pick and rank Curriculum vitaes (Cv) based on job descriptions in huge quantities. Experimental results indicate the performance of the proposed system as an average text parsing accuracy of 85% and a ranking accuracy of 92%.

关键词： Resume ranking Recommender systems Hiring Machine learning Cosine similarity KNN

来源：评论

学校读者我要写书评

暂无评论

Cloud-Driven Collaborative Filtering and Waiter Robots Transforming Customer Experiences in Restaurants

Cloud-Driven Collaborative Filtering and Waiter Robots Trans...

引用

Advances in Modern Age Technologies for Health and engineering science (AMATHE), International Conference on

作者： v. Prasanna srinivasan N. Mohankumar T. Prabakaran A. sairam K. Elangovan s. velmurugan Department of Information Technology R.M.D. Engineering College Chennai Tamil Nadu India Symbiosis Institute of Technology Symbiosis International (Deemed University) Nagpur Maharashtra India Department of Computer Science and Engineering Joginpally B.R. Engineering College Hyderabad Telangana India Department of Computer Science and Engineering Saveetha School of Engineering Saveetha Institute of Medical and Technical Sciences Thandalam Chennai Tamil Nadu India Department of Mechanical Engineering Dhanalakshmi Srinivasan Engineering College Chennai Tamil Nadu India Department of Electronics and Communication Engineering T.J.S. Engineering College Chennai Tamil Nadu India

ISBN: (数字)9798350371567

ISBN: (纸本)9798350371574

To improve customer relations, this research investigates the possibility of using cloud-enabled collaborative filtering with robot waiters in restaurant environments. With the use of innovation, to revolutionize the way people eat out. It suggests using smart robots that can interact with customers and help the wait staff out. Customer preferences are analyzed, order suggestions are streamlined, and service is personalized using cloud-based collaborative filtering algorithms. Test this novel approach in a real-world restaurant setting to view how it changes things for the better in terms of customer happiness, service efficiency, and employee output. It seems that the combination of autonomous waiters and collaborative filtering on the cloud might completely change the way people dine out by making everything more streamlined and personalized to each individual's preferences. To the existing conversation about how robots, cloud computing, and customer service intersect in the restaurant industry, which is important since technology is still playing a major influence on how businesses engage with their customers.

关键词： Industries Training Measurement Cloud computing Technological innovation service robots Robot kinematics

来源：评论

学校读者我要写书评

暂无评论

Enhancing speech Recognition for Hearing-Impaired Individuals Using Deep Learning Models

Enhancing Speech Recognition for Hearing-Impaired Individual...

引用

International Conference on Circuit, Power and Computing Technologies (ICCPCT)

作者： K. Kalaiarasi Winson Medidhi Niresh Kumar s s. Thenappan R. sathiyaseelan Mohana sundaram K Department of ECE Saveetha School of Engineering Saveetha Institute of Medical and Technical Sciences Chennai India Department of CSE(Data Science) Prasad V. Potluri Siddhartha Institute of Technology Kanuru Vijayawada Andhra Pradesh India Department of CSE St. Joseph's College of Engineering OMR Chennai Department of Electronics and Communication Engineering Vel Tech Rangarajan Dr.Sagunthala R&D Institute of Science and Technology Chennai Department of Computer Science Engineering Arunai Engineering College Thiruvannamalai Tamil Nadu India Department of Computer Science and Engineering R.M.D. Engineering College Kavaraipettai Thiruvallur India

ISBN: (数字)9798350372816

ISBN: (纸本)9798350372823

This study examines the constraints of current methodologies, specifically in diverse acoustic environments, to surmount the obstacles associated with speech recognition for individuals with hearing impairments. Conventional processes frequently experience reduced accuracy and utility as a result of their incapability to dynamically adjust to evolving circumstances. This study presents an innovative approach that surpasses these constraints through the utilization of a transformer-based model. Through the implementation of sophisticated data augmentation techniques, the proposed model demonstrates exceptional performance across various acoustic environments, notwithstanding obstacles such as speaker adjustments and background noise. One notable characteristic of this methodology is its transformer architecture, which aptly captures subtle speech inflections by employing self-attention processes. Additionally, through independent learning, the model improves its ability to address the unique challenges associated with hearing-impaired speech. Integrating these components, the model achieves superior accuracy and adaptability compared to current technologies.

关键词： Adaptation models Accuracy speech recognition Auditory system Artificial neural networks speech enhancement Transformers

来源：评论

学校读者我要写书评

暂无评论

An Accurate Prediction and Diagnosis of Alzheimer’s Disease using Deep Learning

An Accurate Prediction and Diagnosis of Alzheimer’s Disease...

引用

North Karnataka subsection Flagship International Conference (NKCon), IEEE

作者： s baskar Nandhini I Mudarakola Lakshmi Prasad Tejaswini Katale Nipun sharma Pundru Chandra shaker Reddy Electrical & Electronics Engineering Dr. MGR Educational and Research Institute Chennai Tamil Nadu India Department of Information Technology V. S. B Engineering College Karur Tamil Nadu India Computer Science and Engineering Institute of Aeronautical Engineering Dundigal Telangana India Department of Computer Engineering Vishwakarma Institute of Technology Pune Maharastra India Electronics and Communication Engineering Presidency University Bangalore Karnataka India School of Computer Science and Engineering Manipal University Jaipur Jaipur Rajasthan India

Recent years have seen an increase in the amount of focus placed on the study of Alzheimer's disease (AD) categorization and prediction. Historically, one would handle such an issue by employing a combination of several types of machine learning algorithms, including regular machine learning algorithms and deep learning algorithms. However, academics are making ongoing efforts to seek for the possibilities of just employing the deep learning method. In order to accurately diagnose this disease, early diagnosis and detection are required. Due to the vast number of people who present themselves with the disease, manual diagnosis by medical professionals is fraught with the possibility of error and is extremely time consuming. The analysis and classification of AD have been approached using a variety of approaches; nonetheless, there is a pressing want for early analysis solutions that are more accurate. The model that was developed as a result of this research provides a deep-learning (DL)-based clarification for the analysis and classification of AD. This solution would make use of the DenseNet-169 &ResNet-50 CNN structures. The AD is categorized in the proposed model as not having dementia, having very mild-dementia, having mild-dementia, and having moderate-dementia. During both the train and the testing stages, the DenseNet-169 demonstrated superior performance. The accuracy values for train &test for DenseNet-169 are respectively 98.7% and 87.36%. In comparison, the accuracy for ResNet-50 were respectively 93.70% and 85.8%. The analysis and classification of AD can be done in real time using the model that was proposed.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Convergence of Data science and blockchain 19

Convergence of Data Science and Blockchain

引用

19th Orissa Information Technology society International Conference on Information Technology, OCIT 2021

作者： shugani, Ashish Namratha, M Kumar, shashank Sabre Travel Technologies Department of Travel Solution Karnataka Bengaluru India B.M.S College of Engineering Department of Computer Science And Engineering Karnataka Bengaluru India

ISBN: (纸本)9781665416641

since the terms 'Data science' and 'blockchain' were coined in 2008 and 1991, both these technologies have seen massive improvement and are being utilized by industries and individuals, across the globe. since their inception, industries have been able to utilize their potential to solve real world problems. The one thing that has been neglected is to find out the areas of convergence of both these technologies to solve a common purpose, that is decentralization and to extract meaningful information and to predict future patterns and behaviors. In this paper, we investigate how blockchain and Data science can converge to complement each other and areas which are difficult to converge so that we are able to harness the full potential of both these technologies for the good of mankind. © 2021 IEEE.

关键词： blockchain

来源：评论

学校读者我要写书评

暂无评论

Risk Modelling and Prediction of Financial Management Using FTL-CsGRU Approach

Risk Modelling and Prediction of Financial Management Using ...

引用

Electronics, Computing, Communication and Control Technology (ICECCC), International Conference on

作者： G. Mothilal Nehru Roopali P. Kudare Amrita Tatia Karnawat Rohini J saurabh singh Neerav Nishant Department of computer Applications Faculty of Science & Humanities SRM UNIVERSITY Chengalpattu India Department of MBA PCET's S B Patil Institute of Management Nigdi Pune Maharashtra India Indira School of Business Studies Tathawade Pune Maharashtra India Computer Science and Engineering Erode Sengunthar Engineering College Perundurai Tamil Nadu India Department of Finance Prestige Institute of Management and Research Indore India Department of Computer Science and Engineering School of Engineering Babu Banarasi Das University Lucknow India

ISBN: (数字)9798350371802

ISBN: (纸本)9798350371819

The need of building information technology is becoming more evident to those working in corporate financial management as the digital era arrives. An increasing number of modern societies are using enterprise resource planning (ERP) systems for their financial management needs. That way, companies can collect financial data in real-time and analyze it later. The steps of preprocessing, feature extraction, and training the model are all sequentially laid out. When it comes to statistical approaches used for data preprocessing, z-score normalization, decimal scaling, and min-max all offer very accurate results. You get the greatest results with MCML when you compare it to NCA, PCA, and Cluster-cluster LDA, the four approaches that comprise feature extraction. The suggested method employed an FTL-CsGRU strategy for model training. This method seems antiquated in comparison to the cutting-edge FTL and GRU. The statistics showed a significant improvement with an accuracy rating of 97.36%.

关键词： Training Accuracy Computational modeling Machine learning Financial management Predictive models Feature extraction

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：