检索结果-内蒙古大学图书馆

Disruptive Technologies (ICDT), International Conference on

作者： Chengamma Chitteti Mallavarapu Sravika Shaik Ayesha Dungavath Sridhar Naik Vangipuram Anil Kumar Muthyamaina Pavan kumar Dept. of Data Science Mohan Babu University (Erst While Sree Vidyaniketan Engineering College) Tirupati A.P India Dept. of CSSE Sree Vidyanikethan Engineering College Tirupati A.P India

ISBN: (数字)9798331519582

ISBN: (纸本)9798331519599

Research shows that fatigue ranks within the three main road accident causes worldwide. This paper develops a driver sleep detection system in real time with MobileNet as the light weight convolutional neural network architecture built specifically for indoor and mobile environments. Real-time system facial scans seek driver indicators of drowsiness which include lengthy eye closures and slow yawn activity leading to sleepiness. MobileNet operates at a low cost which makes it both rapid and functional for vehicle implementation without accuracy compromises. The system helps prevent driver fatigue which leads to better vehicle safety because drivers receive instant feedback through this alert model. This solution presents several advantages such as affordability as well as mobility through mobile devices which makes it ideal for current vehicle safety requirements. Fatigue signs to avoid accidents. When drowsiness accelerates the slowdown of reaction times but increases alertness it causes higher risks of vehicle crashes. System implementations that depend on physical signals need drivers to maintain close positions with the equipment which results in driver distraction. We evaluate the driver eye movement behavior to measure driver behavior through imaging-based dataset measurements in this paper. MobileNet eliminates optical images successfully at a low cost thus speeding up the visual fatigue relief process. MobileNet model demonstrated a performance of 94 % in the evaluation results. The system requires 0.00829 seconds to recognize tiredness. The system delivers exceptional real-time functionality thus making it an extraordinary example. Better performance was achieved through the implementation of hyperparameter optimization in connection with k fold validation. The study matters greatly because its implementation would assist in protecting lives by presenting an effective strategy to stop road fatalities arising from unpredictable seizures.

关键词： Visualization Accuracy Road accidents Costs Sleep Vehicle safety Fatigue Real-time systems Road safety Vehicles

来源：评论

学校读者我要写书评

暂无评论

AutoMediVision:A Deep Learning System for Accurate Pharmaceutical pill identification

AutoMediVision:A Deep Learning System for Accurate Pharmaceu...

引用

data science, Agents & Artificial Intelligence (ICDSAAI), International Conference on

作者： G Rajeswarappa B. Venkata Sivaiah M. Navya Y. Lasya Shaik Mohammad Sameer V. Varun Kumar Department of CSE G Pulla Reddy Engineering College Kurnool Dept. of Data Science School of Computing Mohan Babu University (Erstwhile Sree Vidyanikethan Engineering College) Tirupathi India Dept. of Computer Science and Systems Engineering(CSSE) Sree Vidyanikethan Engineering College Tirupati India

ISBN: (数字)9798331537555

ISBN: (纸本)9798331537562

The goal of this project is to build a high performance system that offers prediction of pill information from images, the pill name, medical use and recommended dosage by using MobileNet which is a lightweight and efficient deep learning model. This system serves as an important tool for healthcare professionals and patients as it offers a means of reducing risk because of the large number of medications out there. For this task, MobileNet was chosen for its high accuracy and good balance between computational efficiency and high accuracy, and hence can be deployed on mobile devices. Having trained the model on a wide variety of pills with class wise annotations, it generalized well for diverse pills. The output of the system is the pill name, primary medical indications and common dosage instructions that help in safer medication management by supporting healthcare providers in providing better care and patient safety.

关键词： Deep learning Accuracy computational modeling Medical services Mobile handsets Safety Standards Research and development Biomedical imaging Pharmaceuticals

来源：评论

学校读者我要写书评

暂无评论

A Machine Learning and data Analytics Approach to Patient Risk Stratification

A Machine Learning and Data Analytics Approach to Patient Ri...

引用

data science, Agents & Artificial Intelligence (ICDSAAI), International Conference on

作者： Veeramalai Sankaradass V.K Manindra Manish Rajesh Anandhan Ramprakash Velmurugan Sabarinathan Sakthivel Dept. of Computer Science and Engineering Chennai Institute of Technology Chennai India Dhanalakshmi Srinivasan Institute of Medical Sciences and Hospital Perambalur India Dept. of Artificial Intelligence and Data Science Chennai Institute of Technology Chennai India

ISBN: (数字)9798331537555

ISBN: (纸本)9798331537562

In this paper, we introduce a robust machine learning (ML) and data analytic solution for risk stratification of patients to promote clinical decision support systems (CDSS) that provide enhanced protection of the patient health interests. Our approach, which we have shown would minimally intrude into peoples' lives, uses predictive models to analyze patient data to identify high-risk patients early in the history of a wide range of medical conditions. Utilizing integrated structured data (e.g., lab results, vital signs) and unstructured data (e.g., clinical notes), our system improves the accuracy of medical decisions made by healthcare providers. We use a combination of supervised learning algorithms and real-time data processing to recalibrate risk scores, which serve as actionable insights for anticipatory interventions. The strategy also comes with extensive validation protocols to verify the efficacy of her model in a healthcare environment. Our results indicate that this practice dramatically increases the accuracy of risk assessments, with important implications for patient care and resource utilization.

关键词： Decision support systems data analysis Accuracy Protocols Supervised learning Real-time systems Risk management Resource management Protection Random forests

来源：评论

学校读者我要写书评

暂无评论

Comprehensive Analysis of Machine Learning Approaches for Phishing Website Detection

Comprehensive Analysis of Machine Learning Approaches for Ph...

引用

Cognitive Computing in engineering, Communications, sciences and Biomedical Health Informatics (IC3ECSBHI), International Conference on

作者： Pragna K P Suhani Shetty Ashwini Kodipalli Ushasree. A Shridhar B. Devamane Roshan Joy Martis Dept. of Electronics and Communication Engineering Global Academy Of Technology Bengaluru Karnataka India Dept. of Artificial Intelligence & Data Science Global Academy of Technology Bengaluru Karnataka India Dept. of Computer Science & Engineering Gokaraju Lailavathi Engineering College Hyderabad India Department of Computer Science School of Computing Amrita Vishwavidyapeetam Mysuru Dept. of Electronics & Communication Engineering Manipal Institute of Technology Bengaluru Manipal Academy of Higher Education Manipal India

ISBN: (数字)9798331518523

ISBN: (纸本)9798331518530

Website phishing, a notable cybercrime, involves deceiving individuals by impersonating legitimate websites. URL spoofing, a common technique in website phishing, involves creating website names. The primary objective of this paper is to accurately differentiate between legitimate and phishing sites using relevant URL attributes and different binary classification machine learning algorithms. This study compares simple, bagging, and boosting algorithms across performance analysis metrics. The best-performing models are selected for an ensemble approach. Additionally, various hyperparameter tuning techniques are assessed for the Random Forest Classifier, and Explainable AI methods like LIME and SHAP are used to interpret findings. The experiment yields optimal parameters, achieving 89. % accuracy with an ensemble model with K-Nearest Neighbors, Decision Tree, XG Boost, and Cat Boost. These findings highlight the significance of ensemble models and hyperparameter tuning in machine learning.

关键词： Uniform resource locators Machine learning algorithms Explainable AI Phishing computational modeling Biological system modeling Nearest neighbor methods Performance analysis Tuning Random forests

来源：评论

学校读者我要写书评

暂无评论

Leveraging Frozen Foundation Models and Multimodal Fusion for BEV Segmentation and Occupancy Prediction

IEEE Open Journal of Vehicular Technology

引用

IEEE Open Journal of Vehicular Technology 2025年 6卷 1241-1261页

作者： Hayes, Seamie Sistu, Ganesh Eising, Ciaran University of Limerick Dept. of Electronic and Computer Engineering Castletroy Co. Limerick V94 T9PX Ireland University of Limerick SFI CRT Foundations in Data Science Castletroy Co. Limerick V94 T9PX Ireland Research Centre Castletroy Co. Limerick V94 T9PX Ireland

In Bird's Eye View perception, significant emphasis is placed on deploying well-performing, convoluted model architectures and leveraging as many sensor modalities as possible to reach maximal performance. This paper investigates whether foundation models and multi-sensor deployments are essential for enhancing BEV perception. We examine the relative importance of advanced feature extraction versus the number of sensor modalities and assess whether foundation models can address feature extraction limitations and reduce the need for extensive training data. Specifically, incorporating the self-supervised DINOv2 for feature extraction and Metric3Dv2 for dept. estimation into the Lift-Splat-Shoot framework results in a 7.4 IoU point increase in vehicle segmentation, representing a relative improvement of 22.4%, while requiring only half the training data and iterations compared to the original model. Furthermore, using Metric3Dv2's dept. maps as a pseudo-LiDAR point cloud within the Simple-BEV model improves IoU by 2.9 points, marking a 6.1% relative increase compared to the Camera-only setup. Finally, we extend the famous Gaussian Splatting BEV perception models, GaussianFormer and GaussianOcc, through multimodal deployment. The addition of LiDAR information in GaussianFormer results in a 9.4-point increase in mIoU, a 48.7% improvement over the Camera-only model, nearing state-of-the-art multimodal performance even with limited LiDAR scans. In the self-supervised GaussianOcc model, incorporating LiDAR leads to a 0.36-point increase in mIoU, representing a 3.6% improvement over the Camera-only model. This limited gain can be attributed to the absence of LiDAR encoding and the self-supervised nature of the model. Overall, our findings highlight the critical role of foundation models and multi-sensor integration in advancing BEV perception. By leveraging sophisticated foundation models and multi-sensor deployment, we can further model performance and reduce data requirem

关键词： Gaussian distribution

来源：评论

学校读者我要写书评

暂无评论

Integrated and Intelligent Elder-Friendly E-commerce Framework using NLP Techniques

Integrated and Intelligent Elder-Friendly E-commerce Framewo...

引用

IEEE Students' Conference on Electrical, Electronics and Computer science (SCEECS)

作者： Sivasanjith G Mohan Prabhu N Hemaswathi S Rajkumar P Rajeshkannan C R. Dhivya Dept. of Information Technology KARE Krishnankovil India Artificial Intelligence and Data Science Indra Ganesan College of Engineering Tiruchirappalli India Dept. of Computer Science and Engg. Mount Zion College of Engg. and Tech. Pudukottai India Dept. of Computer Science and Engg. Suguna College of Engineering Coimbatore India Dept. of Information Technology M.Kumarasamy College of Engineering Karur India

ISBN: (数字)9798331529833

ISBN: (纸本)9798331529840

All over the world, the population of elderly people is increasing a lot. There needs to be special attention to the welfare of the elderly person to make them confident in their independent quality of living. In recent years, researchers have shown much interest in smart home automation for people who live alone without any human assistance. However, the purchase of medical drugs, diapers, etc., outside the house has to be fulfilled by their family members or caretakers. In most cases, it is not feasible, as the family members may stay long away from the elders. Purchasing online is easy; however, analyzing the best product requires knowledge in each e-commerce application for better quality at a low cost. This paper proposes an integrated and intelligent e-commerce framework to avoid the drawback of purchasing items online in a single window using the NLP technique. The proposed framework provides offers, discounts, star ratings, and product reviews by analyzing other customers’ review data and purchase patterns. It provides interactive data analysis at a single window for the ease of elderly persons. The performance of the proposed framework was evaluated by 120 elderly persons using a simple questionnaire with a statistical analysis of a sample purchase.

关键词： Drugs Sentiment analysis Reviews Statistical analysis Urban areas Stars Smart homes Electronic commerce Older adults Recommender systems

来源：评论

学校读者我要写书评

暂无评论

Enhanced Healthcare Through Metaverse Integration: Explainable AI and Blockchain for Trust, Security, and Immersive Experiences

Enhanced Healthcare Through Metaverse Integration: Explainab...

引用

computational Intelligence Approaches and Applications (ICCIAA), International Conference on

作者： Bilal Ibrahim Sowan Hazem Qattous Mohammad Fasha Ahmad Altarawneh Dept. of Bus. Intell. & Data Analytics University of Petra Amman Jordan Dept. of Software Engineering Princess Sumaya University for Technology Amman Jordan Dept. of Computer Science Brunel University of London London UK

ISBN: (数字)9798331523657

ISBN: (纸本)9798331523664

The metaverse, a dynamic blend of virtual and augmented reality, holds extraordinary promise for reshaping the healthcare landscape. By integrating advanced technologies such as virtual reality (VR), artificial intelligence (AI), blockchain, and digital twins, the metaverse introduces transformative opportunities in patient care, medical education, and healthcare operations. These tools foster immersive, personalized, and data-driven environments, significantly improving diagnostic precision, therapeutic strategies, and training methodologies. This paper provides a framework to utilize this potential while navigating key concerns, including systems scale, user customization, and ethical adequacy. In that case, the framework's goal should be eliminating existing limitations and barriers within healthcare systems by utilizing state-of-the-art technologies with real-world applications. The framework involves virtual consultation sites that enable provider-patient interactions, intelligent virtual agents with specific care recommendations, blockchain integration for data integrity, and dynamic VR segments that are functional in delivering healthcare education and training. The study focuses on scenario-based modeling and the use of pilot deployment in various healthcare settings to test for practicability and feasibility. The research also emphasizes that ethical issues and inclusion during the design and implementation of the innovations should also consider society's values regarding the medical field and promote equal access to healthcare. This framework proves a comprehensive blueprint for continuous progress and places the metaverse as one cornerstone in the global healthcare system.

关键词： Training Ethics Technological innovation Solid modeling Metaverse Medical services Blockchains Digital twins Security Medical diagnostic imaging

来源：评论

学校读者我要写书评

暂无评论

Linguistic Entity Masking to Improve Cross-Lingual Representation of Multilingual Language Models for Low-Resource Languages

arXiv

引用

arXiv 2025年

作者： Fernando, Aloka Ranathunga, Surangika Dept. of Computer Science & Engineering University of Moratuwa 10400 Sri Lanka School of Mathematical and Computational Sciences Massey University Palmerston North4443 New Zealand

Multilingual Pre-trained Language models (multiPLMs), trained on the Masked Language Modelling (MLM) objective are commonly being used for cross-lingual tasks such as bitext mining. However, the performance of these models is still suboptimal for low-resource languages (LRLs). To improve the language representation of a given multiPLM, it is possible to further pre-train it. This is known as continual pre-training. Previous research has shown that continual pre-training with MLM and subsequently with Translation Language Modelling (TLM) improves the cross-lingual representation of multiPLMs. However, during masking, both MLM and TLM give equal weight to all tokens in the input sequence, irrespective of the linguistic properties of the tokens. In this paper, we introduce a novel masking strategy, Linguistic Entity Masking (LEM) to be used in the continual pre-training step to further improve the cross-lingual representations of existing multiPLMs. In contrast to MLM and TLM, LEM limits masking to the linguistic entity types nouns, verbs and named entities, which hold a higher prominence in a sentence. Secondly, we limit masking to a single token within the linguistic entity span thus keeping more context, whereas, in MLM and TLM, tokens are masked randomly. We evaluate the effectiveness of LEM using three downstream tasks, namely bitext mining, parallel data curation and code-mixed sentiment analysis using three low-resource language pairs English-Sinhala, English-Tamil, and Sinhala-Tamil. Experiment results show that continually pretraining a multiPLM with LEM outperforms a multiPLM continually pre-trained with MLM+TLM for all three tasks. © 2025, CC BY.

关键词： data curation

来源：评论

学校读者我要写书评

暂无评论

Diversity Seeking Techniques for Red-Teaming Large Language Models

Diversity Seeking Techniques for Red-Teaming Large Language ...

引用

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

作者： Seokhan Lee Bonhwa Ku Hanseok Ko Dept. Computer Science and Engineering Korea University Seoul Republic of Korea Machine Learning Big Data Institute Korea University Seoul Republic of Korea Dept. Electrical and Computer Engineering Korea University Seoul Republic of Korea

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

In this paper, we present new techniques for increasing the diversity of red-teaming prompts generated by automated machine learning-based methods, thereby enabling the discovery of more vulnerabilities in large language models. Using reinforcement learning to train models to output effective prompts for this task results in the models converging deterministically to a single output. Our first technique, which we term Defender, acts by blocking the reward signal for prompts that have already been discovered, thus making what was a stationary problem into a non-stationary problem that compels the reward maximizing algorithm to continually seek new prompts. Our second technique, Teamplay, trains two prompt generation models in tandem and adds the KL divergence between them to the reward in order to make them search in disparate regions of the space of prompts. Our techniques are shown experimentally to increase the effectiveness and diversity of prompts generated by existing reinforcement learning baselines.

关键词： Learning systems Large language models Diversity reception Signal processing algorithms Reinforcement learning Signal processing Acoustics Safety Speech processing

来源：评论

学校读者我要写书评

暂无评论

Explainable Adversarial Attacks on Coarse-to-Fine Classifiers

arXiv

引用

arXiv 2025年

作者： Heidarizadeh, Akram Hatfield, Connor Lazzarotto, Lorenzo Cai, HanQin Atia, George Dept. of Electrical and Computer Engineering University of Central Florida OrlandoFL United States School of Technology Pontifícia Universidade Católica do Rio Grande Sul Porto Alegre Brazil Dept. of Statistics and Data Science Dept. of Computer Science University of Central Florida OrlandoFL United States Dept. of Electrical and Computer Engineering Dept. of Computer Science University of Central Florida OrlandoFL United States

Traditional adversarial attacks typically aim to alter the predicted labels of input images by generating perturbations that are imperceptible to the human eye. However, these approaches often lack explainability. Moreover, most existing work on adversarial attacks focuses on single-stage classifiers, but multi-stage classifiers are largely unexplored. In this paper, we introduce instance-based adversarial attacks for multi-stage classifiers, leveraging Layer-wise Relevance Propagation (LRP), which assigns relevance scores to pixels based on their influence on classification outcomes. Our approach generates explainable adversarial perturbations by utilizing LRP to identify and target key features critical for both coarse and fine-grained classifications. Unlike conventional attacks, our method not only induces misclassification but also enhances the interpretability of the model’s behavior across classification stages, as demonstrated by experimental results. Copyright © 2025, The Authors. All rights reserved.

关键词： Generative adversarial networks

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：