检索结果-内蒙古大学图书馆

27th European Conference on Artificial Intelligence, ECAI 2024

作者： Wang, Boyu Ma, Yue Qiu, Qinru Dept. of Electrical Engineering and Computer Science Syracuse University United States

ISBN: (纸本)9781643685489

Continual learning is referred to as machine learning model's ability to learn a sequence of tasks or data over time without forgetting previously learned knowledge. In particular, we focus on domain incremental learning where the model trained on one domain, e.g., real-life photos, has to be augmented to accommodate data from new domains, e.g., cartoon images. Typical incremental learning relies on rehearsal-based methods, which store trained samples in a buffer and replay them during training alongside new data. This results in significant memory overhead and raise concerns about data privacy. Recently, prompt-based methods address these challenges and outperform them by utilizing the pre-trained Vision Transformer (ViT) and replacing the replay buffer with a prompt pool. However, existing prompt-based models fail to capture domain-specific knowledge and perform poorly in domain incremental learning. In this paper, we propose "domain prompt incremental learning via dynamic neural network", which combines the advantages of architecture-based and prompt-based methods. Specifically, our framework maintains two types of prompt: a instance-level prompt that improves the model's generalization ability is shared across all input samples;and a domain prompt that encodes domain-specific knowledge is assigned for each task. Furthermore, a separated classification head is trained for each domain so that the model has a pre-trained ViT and an ensemble of classification layers, one for each domain. The experimental results shows that our approach outperforms state-of-the-art methods by 2.3% in average on two domain incremental learning (DIL) benchmarks. © 2024 The Authors.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Enhanced Risk Assessment of Human Health Through Gated Dual-Path RNN and Gravitational Search Algorithm 3rd

Enhanced Risk Assessment of Human Health Through Gated Dual-...

引用

3rd International Conference on Emerging Trends in Expert Applications and Security, ICE-TEAS 2024

作者： Shanthi, M.A. Karthik, S. Sabitha, R. Kavitha, M.S. Department of Computer Science and Engineering SNS College of Engineering Tamilnadu Coimbatore641107 India Department of Computer Science and Engineering SNS College of Technology Tamilnadu Coimbatore641035 India

ISBN: (纸本)9789819739905

The need for precise risk assessment of human health in the healthcare industry has resulted in the integration of advanced technology. This study introduces a unique strategy that combines the Gravitational Search Algorithm (GSA) with Gated Dual-Path Recurrent Neural Networks (RNN) to address the present constraints in risk assessment methodologies. Conventional methods are challenged by the complexity of health-related data and the requirement for real-time risk assessment. The goal of this project is to integrate cutting-edge artificial intelligence and optimization techniques to transform the assessment of risks to human health. Predictions produced by current risk assessment algorithms are frequently inadequate due to their inability to identify complex patterns in health data. Moreover, the accuracy and resilience of these models are hampered by the absence of effective optimization tools. For health risk assessment, there is a noticeable lack of integration between advanced neural network topologies and cutting-edge optimization methods. To close this gap and improve risk prediction accuracy, Gated Dual-Path RNN and Gravitational Search Algorithm are combined. The proposed approach uses the Gravitational Search Algorithm to optimize the model parameters and Gated Dual-Path RNN to take advantage of the temporal relationships found in health data. The model capacity to capture both short- and long-term interdependence is improved by the dual-path architecture, and the gravitational search method effectively adjusts the network parameters. Experiments reveal encouraging outcomes, with increased precision and dependability above conventional risk assessment algorithms. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

关键词： Recurrent neural networks

来源：评论

学校读者我要写书评

暂无评论

Breast Cancer Image Classification: Leveraging Deep Learning and Large Language Models for Semantic Integration 2

Breast Cancer Image Classification: Leveraging Deep Learning...

引用

2nd IEEE International Conference on Contemporary Computing and Communications, InC4 2024

作者： Harini, K.K. Nandhini, R. Rajeswari, A.M. Deepalakshmi, R. Velammal College of Engineering and Technology Department of Computer Science and Engineering Madurai India

ISBN: (纸本)9798350383652

Breast cancer, a significant global health problem that is mainly affecting women, demonstrates the importance of early detection to enhance survival rates. Medical image classification is a significant field that utilizes deep learning methods such as Convolutional Neural Networks (CNNs), Region-Based CNNs (R-CNNs), and Deep Neural Networks (DNNs). Despite this, challenges such as over fitting persist, particularly with limited data or complex models. Although these models learn hierarchical features, interpreting them in intricate image structures is still challenging and complicates decision rationale. To address this, we propose a comprehensive approach for breast cancer image classification. The initial process involves image segmentation using CNN to identify regions of interest, followed by precise object detection with OpenCV. Data quality and diversity are ensured by preprocessing and augmentation with ImageDataGenerator, while informative features are extracted using the InceptionV3model. Our analysis of textual descriptions and embeddings is done by integrating Sentence Transformers and CLIP from the Large Language Model (LLM) framework. Achieving robust decision-making is ensured by setting a threshold of 0.7 for image classification based on predicted malignancy probability. Finally, improving the computations of similarity scores can improve prediction accuracy for reliable diagnostic interpretations. An accuracy of 0.92 in breast cancer diagnosis is guaranteed by integrating LLM and CNN for 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

On Using BERT Embeddings for Text Semantic Communication 21

On Using BERT Embeddings for Text Semantic Communication

引用

21st IEEE International Conference on Smart Communities: Improving Quality of Life using AI, Robotics and IoT, HONET 2024

作者： Mahgoub, Asma Yaacoub, Elias College of Engineering Qatar University Department of Computer Science and Engineering Doha Qatar

ISBN: (纸本)9798350378078

The current technological advancements carry stringent communication requirements. These requirements cannot be achieved by current networks. Therefore, a new paradigm called semantic communication is proposed. Semantic communication is the transmission of the meaning of data. Different semantic communication algorithms were proposed in literature, but they are complex and non-universal. In this work, a text semantic communication algorithm will be modified to be simple and universal. The modified algorithm uses a transformer network with 1 encoder layer and 3 decoder layers. The embedding layer uses pretrained embeddings from a Bidirectional Encoder Representations from Transformers (BERT) model. The algorithm is trained to optimize both the cross-entropy loss and the mutual information. Training and testing are done using the European parliament proceedings dataset. The performance of the modified algorithm is comparable to the original algorithm although it uses 50% of the layers. This work indicates the potential of using pretrained embeddings to simplify and enhance universality of existing semantic communication algorithms. © 2024 IEEE.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Bidirectional Sign Language Translation System 8

Bidirectional Sign Language Translation System

引用

8th IEEE International Conference on Computing, Communication, Control and Automation, ICCUBEA 2024

作者： Mahjan, Aditya Hendre, Pradnya Mankhair, Gauri Satpute, Shubhangi Chaudhari, Deepti A. Nutan College of Engineering and Research Department of Computer Science and Engineering Pune India

ISBN: (纸本)9798350391770

Sign language is an essential form of communication for the deaf and hard-of-hearing community. The creation of Bidirectional Sign Language Translation Systems (BSLTS) has drawn a lot of attention as technology has advanced. The first section of the article gives a general review of the difficulties in translating and recognising sign language, highlighting the urgent need for precise and timely solutions. After that, it explores the fundamental ideas of CNN and Sign language, emphasising their advantages. The survey discusses different architectures and techniques used for feature extraction and classification, and it thoroughly evaluates the body of research on CNN's use in sign language recognition. Additionally, the study investigates how the cutting-edge object detection technology YOLO3 improves the spatial comprehension of sign language gestures. This investigation investigates how real-time object identification capabilities of YOLO3 can enhance the accuracy and efficiency of BSLTS, enabling smooth communication between sign language users and non-signers. Sign language is a communication method used by people with speech and hearing impairments to interact with others who are not handicapped. Normal people don't use sign language very often, and it takes a lot of time and effort to understand signs. People with hearing impairments are cut off from the outside world and seldom engage in social interactions due to this communication gap. A system for converting text to sign and sign to text is required to tackle this issue and facilitate communication. For system training, a variety of methods have been employed, including Sign language and Convolutional Neural Networks (CNN). © 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

Targeted materials discovery using Bayesian algorithm execution

引用

npj Computational Materials 2024年第1期10卷 1619-1630页

作者： Sathya R.Chitturi Akash Ramdas Yue Wu Brian Rohr Stefano Ermon Jennifer Dionne Felipe Hda Jornada Mike Dunne Christopher Tassone Willie Neiswanger Daniel Ratner SLAC National Accelerator Laboratory Menlo ParkCA94025USA Department of Materials Science and Engineering Stanford UniversityStanfordCA94305USA Department of Computer Science Stanford UniversityStanfordCA94305USA Department of Computer Science University of Southern CaliforniaLos AngelesCA90089USA

Rapid discovery and synthesis of future materials requires intelligent data acquisition strategies to navigate large design spaces.A popular strategy is Bayesian optimization,which aims to find candidates that maximize material properties;however,materials design often requires finding specific subsets of the design space which meet more complex or specialized *** present a framework that captures experimental goals through straightforward user-defined filtering *** algorithms are automatically translated into one of three intelligent,parameter-free,sequential data collection strategies(SwitchBAX,InfoBAX,and MeanBAX),bypassing the time-consuming and difficult process of task-specific acquisition function *** framework is tailored for typical discrete search spaces involving multiple measured physical properties and short time-horizon decision *** demonstrate this approach on datasets for TiO2 nanoparticle synthesis and magnetic materials characterization,and show that our methods are significantly more efficient than state-of-the-art ***,our framework provides a practical solution for navigating the complexities of materials design,and helps lay groundwork for the accelerated development of advanced materials.

关键词： synthesis straight consuming

来源：评论

学校读者我要写书评

暂无评论

A Data-Driven Approach for Online Phishing Activity Detection

A Data-Driven Approach for Online Phishing Activity Detectio...

引用

2024 International Conference on Advances in Computing, Communication, electrical, and Smart Systems, iCACCESS 2024

作者： Ahmad, Kazi Injimam Paul, Anirban Hafiz, Md. Ferdous Bin Chakraborty, Sovon Khan, Niaz Ashraf Dipto, Shakib Mahmud East Delta University Department of Computer Science and Engineering Chittagong Bangladesh University of Liberal Arts Bangladesh Department of Computer Science and Engineering Dhaka Bangladesh

ISBN: (纸本)9798350350289

The exponential rise in online activity has increased the possibility of phishing attempts, which can seriously jeopardize sensitive data and put companies and individuals at risk. This study uses cutting-edge machine learning and deep learning approaches to solve the urgent need to identify phishing websites efficiently. The proposed research utilizes two different extensive datasets with 30 and 87 features, respectively, that have been extracted from various network traffic, site content, and URLs to broadly measure the binary classification performance and accuracy of 5 different machine learning models as well as 3 deep learning-based models. Furthermore, this research employed one of the most popular feature selection techniques, Pearson correlation, to reduce the number of input variables, minimizing the amount of redundant or irrelevant features in the datasets. Training our machine learning and deep learning models with the resulting mostly redundancy-free subset of features was very efficient and with the addition of hyperparameter tuning and 10-fold cross-validation, our models produced some of the most accurate prediction results in web phishing detection. Among all the tests, the Random Forest algorithm performed better in the limited-sized dataset and scored a prediction accuracy of 97.83%. On the larger dataset, eXtreme Gradient Boosting performed slightly better and achieved a prediction accuracy of 97.02%. © 2024 IEEE.

关键词： Classification (of information)

来源：评论

学校读者我要写书评

暂无评论

YOLOV7-BASED HELMET DETECTION SYSTEM FOR TRAFFIC SAFETY 6

YOLOV7-BASED HELMET DETECTION SYSTEM FOR TRAFFIC SAFETY

引用

6th International Conference on Computational Intelligence and Communication Technologies, CCICT 2024

作者： Malik, Indu Verma, Satish Sachan, Sarthak Rohit Department of Computer Science and Engineering ABES Engineering College UP Ghaziabad India

ISBN: (纸本)9798350374629

The current predominant issue in India is the escalating number of traffic accidents, primarily attributed to the burgeoning population. A significant contributing factor to this problem is the widespread disregard for helmet usage, leading to an increase in accidents. This project aims to address this issue by implementing a system capable of recognizing individuals who are not wearing helmets and subsequently advising them to prioritize helmet use in the future. The fundamental principle guiding this system is deep learning, with the project encompassing two key phases: the initial detection of individuals and the subsequent identification of helmets using YOLOv7. The utilization of the YOLOv7 algorithm, launched in 2022, is specifically chosen for its accuracy and favorable results. In addressing the associated challenges, this project introduces a Non-Helmet Rider Detection system. This system strives to automate the identification of traffic violations related to non-helmet usage and extract license plate numbers from vehicles. Object detection through deep learning is implemented at three levels within this system. The primary objective is to mitigate the occurrence of head injuries by ensuring that individuals in potentially perilous environments adhere to the requisite protective headgear. ©2024 IEEE.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Self-annotated Labelling and Training Data for Traffic Video Object Detection Using Machine Learning Techniques 7th

Self-annotated Labelling and Training Data for Traffic Video...

引用

7th IFIP TC 12 International Conference on Computational Intelligence in Data science, ICCIDS 2024

作者： Rahul Chiranjeevi, V. Swamy, Maanesh M. Krishna Prasath, M.K. Kumar, P. Department of Computer Science and Engineering Rajalakshmi Engineering College Thandalam Chennai India

ISBN: (纸本)9783031699818

Classical machine learning algorithms are susceptible to objective elements like video quality and the weather, which results in inferior detection results in an erroneous identification of Unmanned Aerial Vehicle (UAV) action photos. Deep learning is suggested for identifying cars in UAV video. Thanks to the technology offered by road management systems, real-time visual information is now accessible in thousands of locations on road networks. The first step in identifying or preventing accidents is locating the vehicles on the path of travel. Convolutional neural networks have made substantial advancements in the field of object detection. Enhancing conventional computer vision techniques. However, there are drawbacks because the pre-trained models that are currently available only provide a low detection rate, especially for small objects. The major drawback is that, in order to retrain the automobile maker, they must manually identify the vehicles that appear in the photographs captured by each IP camera on the road infrastructure. This task will be impossible to complete even if we install hundreds of cameras throughout the extensive road network. The technique utilized in this study uses CCTV footage with a comparable range of images with the aim of identifying cars in an image. The spatial brightness of the original sample is first translated using the HSV (Hue, Saturation, Value) method to maximize sample diversity and adaptability to different lighting conditions. The performance of the basic SSD is enhanced by incorporating focus loss because of feature extraction. Following a trained neural network model’s examination of the UAV footage, the effectiveness of drone identification is assessed. The results show that the proposed method have a car detection rate of 96.49%. © IFIP International Federation for Information Processing 2024.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

DSTER: A Dual-Stream Transformer-based Emotion Recognition Model through Keystrokes Dynamics 18

DSTER: A Dual-Stream Transformer-based Emotion Recognition M...

引用

18th IEEE International Joint Conference on Biometrics, IJCB 2024

作者： Chen, Frank Rao, Shruti Tiwari, Brijesh Phoha, Vir V. Syracuse University Syracuse Department of Electrical Engineering and Computer Science New York13210 United States

ISBN: (纸本)9798350364132

Emotion Recognition is a critical research area for enhancing human-computer interaction. Keystroke dynamics, a behavioral biometric capturing typing patterns, offers a non-intrusive, user-friendly method for recognizing emotions. We propose a Dual-Stream Transformer-based Emotion Recognition (DSTER) model, which leverages keystroke dynamics to determine emotional states. The DSTER model features a dual-stream architecture that separately extracts temporal-over-channel and channel-over-temporal information. Each stream employs multi-head self-attention mechanisms, Long-Short Term Memory (LSTM), and Convolutional Neural Network (CNN) layers, along with dense vector embeddings of keycode data, to improve the extraction of temporal and contextual information from typing sequences. To the best of our knowledge, the DSTER model is the first to integrate transformer architecture with keystroke dynamics for emotion recognition. Our experiments on a widely-used fixed-text dataset demonstrate that the DSTER model significantly outperforms the three most recent baseline models, achieving average F1 scores up to 0.989 and an average accuracy increase of up to 66.04%. Unlike the significant performance variations reported in baseline models, the DSTER model maintains consistent and robust performance across all five tested emotional states. Further analysis shows that the model performs better with longer window lengths and greater overlaps. © 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

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：