检索结果-内蒙古大学图书馆

Collaborative, Privacy-Preserving Federated Learning Framework for the Detection of Diabetic Eye Diseases

SN computer science 2024年第8期5卷 1100页

作者： Gulati, Seema Guleria, Kalpna Goyal, Nitin Chitkara University Institute of Engineering and Technology Chitkara University Punjab Rajpura India Department of Computer Science and Engineering School of Engineering and Technology Central University of Haryana Haryana Mahendergarh India

Diabetes Mellitus is one of the chronic metabolic disorders that is easily identified by elevation in blood sugar levels. The high blood sugar levels over time can lead to substantial damage to the blood vessels, heart, eyes, kidneys, and nerves. Diabetic Eye Diseases (DED) such as Diabetic Macular Edema (DME), Diabetic Retinopathy (DR), and glaucoma, are serious complications of diabetes that can result in irreversible vision loss. Thus, affecting the health and quality of life of the population worldwide. The primary objective of the proposed work is to evaluate the efficacy of a privacy-preserved Federated Learning (FL) framework for detecting DEDs. To detect DEDs a collaborative and privacy-preserving Federated Deep Learning (FDL) framework using a lightweight MobileNetV2 architecture is proposed. The aggregation scheme selected for the FDL framework is FedProx and to achieve a high accuracy score the proximal term, µ is tuned with values ranging from µ= (0.01, 0.02, 0.03, 0.04 and 0.05). The impact of the variance in the proximal term on model performance is analyzed, and the results show that higher µ (0.03,0.4 and 0.5) values provide more stable and consistently high precision, recall, and F1-scores. The findings suggest that FedProx aggregation successfully stabilizes the performance of the framework with the least standard deviation of 0.035 in accuracy at µ = 0.05. The balance between accuracy and stability is a key factor in the successful application of federated learning in disease prediction. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2024.

关键词： Artificial intelligence Collaborative training Diabetic eye diseases Federated learning Privacy-protection

来源：评论

学校读者我要写书评

暂无评论

Multi lingual Video Summarizer using Natural Language Processing 5

Multi lingual Video Summarizer using Natural Language Proces...

引用

5th IEEE International Conference for Emerging technology, INCET 2024

作者： Adithya, T. Sai Divya, P. Vamshiteja, N. Prasad, Andraju Bhanu Hariharan, Shanmugasundaram Kukreja, Vinay Department of Computer Science and Engineering Vardhaman College of Engineering Hyderabad India Chitkara University Institute of Engineering and Technology Punjab India

ISBN: (纸本)9798350361155

Throughout the day, a huge quantity of video recordings is made and shared online. Because most films are uploaded with little to no context, it has become quite difficult to spend time watching these kinds of movies in today's hectic lifestyle. Our efforts may also be in vain if we are unable to find any useful information from them. The quantity of videos that need to be made could be directly impacted by the sharp rise in YouTube users' year over year. The likelihood that the makers of the videos will provide inaccurate information about the original content of the film is certainly very high due to their desperation for views. The user's precious time and resources could be wasted in this way. A summary button is one of the user-friendly features of the summarizer Chrome extension, which is designed to further enhance user interaction. This integration of Gensim algorithm with the summarizer Chrome extension streamlines the browsing experience, helping users navigate the overwhelming amount of online video content effectively. When you click this button, the Google Chrome web browser's Chrome extension shows the synopsis of the YouTube video that is currently playing. © 2024 IEEE.

关键词： Video recording

来源：评论

学校读者我要写书评

暂无评论

A Comprehensive Analysis of Machine Learning and Deep Learning Based Product Recommendation System 5th

A Comprehensive Analysis of Machine Learning and Deep Learni...

引用

5th EAI International Conference on Cognitive Computing and Cyber Physical Systems, IC4S 2024

作者： Chakravarthi, GKalyan Kumar, Raghvendra Addagarla, Ssvr Kumar Department of Computer Science and Engineering School of Engineering and Technology GIET University Odisha Gunupur India Nadimpalli Satyanarayana Raju Institute of Technology Visakhapatnam India

ISBN: (纸本)9783031770746

Recommender Systems (RS) have been widely applied in various real-time applications to support identifying valuable information. The RS tries to give actual suggestions to every user based on their behavior as well as interests. Recommendations generated by these systems often contend with the unique personal interests of individual users, thereby playing a pivotal role in their decision-making processes. Recommender Systems (RS) act as efficient tools for filtering vast amounts of online data, shaping the behaviors of smartphone users, personalization trends, and the evolution of internet access. RS are generally classified into three primary types: Hybrid Filtering, Content-Based Filtering (CBF), and Collaborative Filtering (CF). These systems find wide application across various fields including online education systems, e-commerce, marketing, tourism, food service, movies, business, and beyond. Although the recent RS are well-known in providing valuable recommendations, they suffer from a number of limitations as well as challenges such as scalability, sparsity, and cold-start and so on. Due to the various approach’s existence, the selection of these approaches becomes challenging during the development of application-based RS. Moreover, every approach comes with its individual feature sets, advantages as well as limitations, which must be addressed. This survey reviews the research inclinations which integrates the progressive technical characteristics of the RS. The aim of this survey is to ensure a systematic review on recent contributions in RS domain, and concentrated on various applications such as education, food, products and so on. This survey provides a comprehensive review of these types of RS, recent literature, and applications of visual RS. © ICST institute for computer sciences, Social Informatics and Telecommunications engineering 2025.

关键词： Collaborative filtering

来源：评论

学校读者我要写书评

暂无评论

Automated Exoskeleton Prosthetic Suit With Monitoring System 4

Automated Exoskeleton Prosthetic Suit With Monitoring System

引用

4th International Conference on Power, Energy, Control and Transmission Systems, ICPECTS 2024

作者： Jeeva, C. Sagar, Lalit Kumar Srivastava, Mudit P. Abinaya, R. Narmatha, M. Shanmuga Priya, M. Sri Sairam Engineering College Department of Electrical and Electronic Engineering Tamil Nadu India Srm Institute of Science and Technology Department of Computer Science Engineering Delhi India Srm Institute of Science and Technology Department of Physics Delhi India

ISBN: (纸本)9798331508845

The Automated Exoskeleton Prosthetic Suit with Monitoring System aims to develop a fully functional prosthetic suit with coordinated strength and response time. The suit is ameliorative and can be worn or removed as needed. Designed as a wearable inner suit, it operates autonomously, incorporating sensors and navigation systems to reduce inequality. In regions where the suit is worn, it analyzes blood flow using ultrasound Doppler principles and aids in improving circulation. The exoskeletons support or replace impaired body parts, allowing users to function properly. Electrical motors and precise position control form the core of its most common design. This suit explores the entire body using various actuators, tasks, and degrees of freedom. The product can be worn as an inner suit, featuring an autonomous sensor and navigation system to eliminate inequality. It measures blood flow in the regions covered by the suit, enhancing circulation. The primary goal of the product is to detect even small objects and prevent the wearer from falling. © 2024 IEEE.

关键词： Exoskeleton (Robotics)

来源：评论

学校读者我要写书评

暂无评论

A Novel Approach on Deep Reinforcement Learning for Improved Throughput in Power-Restricted IoT Networks 1

引用

International Conference on Electrical and Electronics engineering, ICEEE 2023

作者： Sweety Bakyarani, E. Singh, Navneet Pratap Shekhawat, Jyoti Bhardwaj, Saurabh Chaku, Shweta Singh, Jagendra Department of Computer Science SRM Institute of Science and Technology Kattankulathur India School of Computer Science Engineering and Technology Bennett University Greater Noida India Vivekanand Global University Rajasthan Jaipur India Department of Information Technology Raj Kumar Goel Institute of Technology AKTU Lucknow Ghaziabad India Department of Computer Science Engineering Inderprastha Engineering College AKTU Lucknow Ghaziabad India

ISBN: (数字)9789819986613

ISBN: (纸本)9789819986606

The rapid expansion of the Internet of Things (IoT) has stressed the importance of energy-efficient communication protocols, particularly in networks operating under power constraints. This paper presents a unique approach for managing communication in energy-limited IoT networks using a deep reinforcement learning (DRL)-based communication protocol. By integrating Sparse Code Multiple Access (SCMA), Code Division Multiple Access (CDMA) techniques, and the Combined Experience Replay Deep Deterministic Policy Gradient (CER-DDPG) algorithm, we developed a novel protocol to improve the throughput of power-constrained sensors in an IoT network. Through comprehensive simulations, we compared the proposed protocol’s performance with benchmark systems like traditional DDPG and stochastic algorithms. The results reveal superior energy efficiency and throughput with the proposed protocol, establishing its potential to significantly enhance the performance of energy-constrained IoT networks. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Detection of Twitter Spam Using GLoVe Vocabulary Features, Bidirectional LSTM and Convolution Neural Network

引用

SN computer science 2024年第2期5卷 206页

作者： Manasa, Pinnapureddy Malik, Arun Batra, Isha Department of Computer Science and Engineering Lovely Professional University Phagwara India Neil Gogte Institute of Technology Hyderabad India

Twitter spam is used to describe any form of unwanted or unsolicited communications, tweets, or activities that users encounter on the social media platform Twitter. This can include things like spam messages from fake accounts, tweets with links to malicious or scam websites, or automated tweets that are sent out in large quantities. Training a model to detect patterns in tweets indicative of spam. The process typically involves collecting a large dataset of labeled tweets, and then using this dataset to train a machine learning model. The proposed model is divided into two parts that operate on the input dataset, which includes tweets and additional meta-data such as the number of followers and actions. The first part focuses on the text of the tweets, using the GLoVe language model to extract vocabulary features which are then used to detect spam using an LSTM deep learning model. The second part utilizes meta-data from the tweets and additional meta-heuristic features, such as the tweet length and number of question marks, and classifies them using a CNN model. The final decision is determined by combining the results from both the LSTM and CNN models. © 2024, The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd.

关键词： CNN GLoVe embeddings LSTM Twitter spam Vocabulary features

来源：评论

学校读者我要写书评

暂无评论

Enhancing Cyberbullying Detection on Twitter with Psychological Features and Machine Learning 1

Enhancing Cyberbullying Detection on Twitter with Psychologi...

引用

1st IEEE International Conference on Emerging Research in Computational science, ICERCS 2023

作者： Lalitha, N. Sumaya Thabasum, S.K. Tejaswini, N. Bhagya Sri, D. Srivani, R. Vignan Nirula Institute of Technology and Science for Women Computer Science & Engineering Guntur India

ISBN: (纸本)9798350359756

Today, a large number of people dabble in the realm of social media. Due to the pandemic situation, people are even more engaged since they frequently use social media to vent their emotions. One of the many detrimental effects of this pervasive usage of social media is cyberbullying, which is a troubling form of online harassment. Though it can take several forms, the most common one is text. Cyberbullying is common on social media, and instead of confronting the perpetrator, victims often have mental breakdowns as a result of it. This study's computerized cyberbullying detection method accesses Twitter users' psychological traits, including their personalities, moods, and emotions. Our study provides an innovative solution for detecting cyberbullying tweets by an attention-based transformer algorithm combined with embeddings. Our model acts as a detector in classifying tweets that are related to cyberbullied actions. These tweets are converted into numerical vectors by Embeddings and divided into fixed segments through the padding technique. The transformer model learns from the encoder part comprising of self-attention and feed-forward neural network and normalization through the tweet's dataset. Incredibly accurate cyberbullying detection is made possible by this integrated technology. Our approach promises to identify cyberbullying quickly and precisely to give more control to women over the situation. © 2023 IEEE.

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

A Novel Approach to Optimise Credit Risk Prediction with Enhanced SVM and and Hinge Loss 1

A Novel Approach to Optimise Credit Risk Prediction with Enh...

引用

1st International Conference on Computational Intelligence for Information, Security and Communication Applications, CIISCA 2023

作者： Dhawan, Kshitij Jayakumar, Naveenkumar School of Computer Science and Engineering Vellore Institute of Technology Vellore India

ISBN: (纸本)9798350339727

When a borrower doesn't keep their end of the bargain, there is a danger of financial loss known as credit risk. Even though there are numerous factors that contribute to credit risk, careful consideration when approving a loan may lessen the likelihood of the development of non-performing assets (NPA) and fraud. While there are organizations that offer credit rating and credit scoring services to banks for a fee, the researchers continue to investigate different machine learning strategies to increase the accuracy of credit risk assessment. In this study, a novel approach based on a modified version of SVM (Support Vector Machines) is proposed. The proposed methodology incorporates a custom-trained variant of SVM. A function calculates class weights based on the inverse frequency of each class in the training set. The hinge loss is calculated using a function that takes the feature matrix, target labels, weight matrix, and regularization parameter as input. The hinge loss function penalizes the model for misclassifying instances. Further, this work compares three machine learning (ML) algorithms-SVM, XGBoost, and Random Forest with the proposed approach to comprehend the efficacy in spotting credit risk. The proposed custom SVM model is trained with new weights that were found using the current dataset, which provided an accuracy of 96.92% that is higher than the pre-existing models. Also, the findings showed an fl-score of 0.90, a precision value of 0.915, and a recall value of 0.885. © 2023 IEEE.

关键词： Support vector machines

来源：评论

学校读者我要写书评

暂无评论

Automatic Detection of Plasmodium Parasite Using Convolution Neural Network

Automatic Detection of Plasmodium Parasite Using Convolution...

引用

2023 International Conference on Self Sustainable Artificial Intelligence Systems, ICSSAS 2023

作者： Janardhan, G. Shriya, D. Babu, D. Akshay Kumar, S. Anjan Kishore, R. Sai Vignan Institute of Technology and Science Computer Science & Engineering Telangana State Hyderabad India

ISBN: (纸本)9798350300857

The health industry is quite active and malaria is the deadliest disease on earth. A laboratory or skilled technician will often do a schematic analysis of a patient's blood film under a microscope to look for parasite-infected red blood cells. The procedure is ineffective, and the diagnosis is based on the expertise and experience of the individuals who will be performing the examination. Malaria diagnosis using a blood smear has already been carried out using deep learning algorithms. The suggested method uses image-based diagnosis with CNNs and VGG models to accomplish this goal, yielding a high accuracy of 94.7% and precision in differentiating infected and uninfected cells. © 2023 IEEE.

关键词： Diseases

来源：评论

学校读者我要写书评

暂无评论

Radish Plant Growth Monitoring Using Multimodal Fusion 1

Radish Plant Growth Monitoring Using Multimodal Fusion

引用

1st IEEE Conference on AgriFood Electronics, CAFE 2023

作者： Bansal, Shreya Singh, Malya Barda, Seema Kumar, Vikas Goel, Neeraj Saini, Mukesh Computer Science & Engineering Indian Institute of Technology Punjab Ropar India

ISBN: (纸本)9798350327113

In the realm of precision agriculture, successful plant growth monitoring plays a pivotal role, and it is greatly aided by advanced deep learning techniques utilizing data from various sensors. The key objective of the paper is to develop a model to monitor the growth of the radish plant through the integration of multi-modal data, specifically RGB and Depth (RGB-D) images. Plant growth can be modeled based on many parameters, for example, the height of the plant, the environmental conditions (such as temperature, humidity, and light intensity), nutrient levels in the soil, leaf count, and the presence of pests or diseases, etc. The proposed study estimates the age of the radish plant with an end-to-end deep-learning model. We developed a Fused Image Transformer (FIT) model to estimate the age of the plant in weeks. The FIT uses a self-attention mechanism to estimate the essential features of the images which will help to complete the objective. The proposed method is compared with the state-of-the-art models for regression problems. We have also validated the FIT model with or without depth information. The Mean squared loss with and without depth information is found to be 0.025 weeks and 2.059 weeks respectively which shows a significant improvement by using the depth information. © 2023 IEEE.

关键词： Deep learning

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：