检索结果-内蒙古大学图书馆

9th IEEE International Conference on Progress in Informatics and Computing, PIC 2024

作者： Yang, Ke Feng, Jingyun Xi, Zhiyu Zheng, Qiuke Zhang, Huan Deng, Wei China Railway Eryuan Engineering Group Co. Ltd. Chengdu China Southern University of Science and Technology Dept. of Statistics Shenzhen China School of Automation Science and Electrical Engineering Beihang University Beijing China School of Computer Science and Artificial Intelligence Southwestern University of Finance and Economics Chengdu China School of Statistics Southwestern University of Finance and Economics Chengdu China Chongqing University of Posts and Telecommunications Chongqing Key Laboratory of Computational Intelligence Chongqing China

ISBN: (纸本)9798350363203

The structure of knowledge graph is basically flat and lack of hierarchies to make knowledge graph too general to deal with concrete reasoning problems. There are few work about hierarchical knowledge graph. In this paper, the entity and relation combined method is proposed to divide the entities into groups based on structure similarity. The hierarchical cluster method is proposed with the entity and relation combined embedding representation to construct a multi-granular knowledge graph. The experimental results show that the hierarchical division relied on combined embedding are significantly better than those based on the existing entity embedding methods in three evaluation indexes. The combined embedding can capture the structure information which is helpful to hierarchical knowledge graph accurately and reasonably. © 2024 IEEE.

关键词： Knowledge graph

来源：评论

学校读者我要写书评

暂无评论

Real-Time Traffic Intensity Estimation and Management using Deep Learning Techniques

Real-Time Traffic Intensity Estimation and Management using ...

引用

Network, Multimedia and Information Technology (NMITCON) International Conference on

作者： Suryakant Murhekar Yash Kamble Snehlata Dongre Priya Dasarwar Tabassum H Khan Vaishali Patil Data Science (Center Of Excellence) G H Raisoni College Of Engineering Nagpur India Dept. of Computer Science and Engineering G H Raisoni College Of Engineering Nagpur India Dept. of Computer Science and Engineering Symbiosis Institute of Technology Nagpur Campus Symbiosis International (Deemed University) Pune India Dept. of Artificial Intelligence G H Raisoni Institute Of Engineering and Technology Nagpur India

ISBN: (数字)9798350372892

ISBN: (纸本)9798350372908

In India, traffic flow estimation is crucial for effective transportation management and urban planning. This research proposes a comprehensive approach leveraging deep learning techniques, particularly the YOLOv8 object detection model, to estimate road traffic flow in urban environments. The model is trained on the widely used COCO dataset, ensuring robust performance in detecting various objects, including vehicles, pedestrians, and cyclists. Subsequently, the model is fine-tuned on a custom dataset comprising annotated traffic images, enabling it to adapt specifically to the complexities of urban traffic scenarios. The research methodology encompasses several key steps. Firstly, the pretrained YOLOv8 model is utilized to perform inference on sample images, demonstrating its capability to accurately detect vehicles within the urban environment. Visualization and analysis of the detection results showcase the model’s effectiveness in identifying vehicles with high confidence levels, laying a foundation for further experimentation. Following the initial assessment, the model undergoes fine-tuning on the custom dataset, optimizing its performance for the task of traffic flow estimation. Training parameters such as epochs, image size, and learning rate are carefully tuned to achieve optimal results. Additionally, metrics such as box loss, classification loss, and distribution focal loss are monitored during training to assess model convergence and performance. Upon completing the training phase, the model’s performance is evaluated using a validation set, providing insights into its generalization ability and effectiveness in real-world scenarios. The evaluation metrics, including precision, recall, and F1 score, offer quantitative measures of the model’s performance, further validating its utility for traffic flow estimation tasks. Furthermore, the research extends beyond static image analysis to real-time traffic intensity estimation using video data. By integratin

关键词： Training Deep learning Adaptation models Analytical models Urban planning Pipelines Estimation Transportation Streaming media Real-time systems

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

Detection of Pre Shoplifting Suspicious Behavior Using Deep Learning 16

Detection of Pre Shoplifting Suspicious Behavior Using Deep ...

引用

16th IIAI International Congress on Advanced Applied Informatics, IIAI-AAI 2024

作者： Shrestha, Sujan Taniguchi, Yoji Tanaka, Tetsuo Graduate school of Engineering Kanagawa Institute of Tehnology Course of Information and computer sciences Atsugi Japan Daiichi Institute of Technology Artificial Intelligence and Data Science Department of Information Tokyo Japan Kanagawa Institute of Technology Faculty of Information Technology Dept. of Information and Computer Sciences Atsugi Japan

ISBN: (纸本)9798350377903

Shoplifting presents a significant challenge for retailers, leading to substantial financial losses and security concerns. Current surveillance systems effectively identify shoplifting incidents postoccurrence;however, the capability to detect and prevent such activities preemptively remains limited. This research focuses on identifying behaviors that likely precede shoplifting events. To this end, we introduce a novel benchmark dataset, specifically designed for the early detection of suspicious behavior that may lead to shoplifting. This dataset is curated by selecting and segmenting relevant sequences from the UCF101-Crime dataset, encompassing suspicious behavior, actual shoplifting incidents, and normal activities. Our approach addresses the limitations of existing datasets where the brief duration of shoplifting events may skew neural network performance. The dataset includes clips of only a few seconds each, for which we propose a robust neural network architecture combining the strengths of convolutional neural networks (CNN) for efficient feature extraction and Bidirectional long short-term memory (BiLSTM) networks for capturing temporal dependencies within action sequences. Our model achieves an impressive classification accuracy of 86.21 % on the newly created dataset. © 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

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

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

In silico drug delivery modeling of a novel Everolimus Coated Balloon 46

In silico drug delivery modeling of a novel Everolimus Coate...

引用

46th Annual International Conference of the IEEE engineering in Medicine and Biology Society, EMBC 2024

作者： Kotoulas, Grigorios G. Loukas, Vasileios S. Karanasiou, Georgia S. Sakellarios, Antonis I. Pezoulas, Vasileios C. Katsouras, Christos C. Michalis, Lambros K. Semertzioglou, Arsen Moulas, Anargyros Geris, Liesbet Fotiadis, Dimitrios I. University of Patras Laboratory of Biomechanics and Biomedical Engineering Department of Mechanical and Aeronautics Engineering Patras Greece University of Ioannina Medical School Dept. of Cardiology GR Ioannina Greece Rontis Corporation S.A Larisa Greece University of Thessaly General Department Larisa Greece University of Liège Ku Leuven Department of Biomechanics and Computational Tissue Engineering Belgium University of Ioannina Unit of Medical Technology and Intelligent Information Systems Dept. of Materials Science and Engineering Ioannina Greece Biomedical Research Institute Foundation for Research and Technology Hellas FORTH-BRI Ioannina Greece

ISBN: (纸本)9798350371499

This study examines the drug elution process of a novel Everolimus Coated Balloon (ECB) that features a novel polymer mixture consisting of cellulose derivatives and accelerators, and Everolimus as the pharmaceutical component, based on commercially available PTCA balloon-catheter system. In this analysis, we utilized in silico postdeployment results comprising three distinct datasets of healthy porcine arteries undergoing PTCA with (a) Sirolimus, (b) Low-dose Everolimus, and (c) High-dose Everolimus. We developed a computational model that consists of the drug release from the balloon surface and the drug transport through the porous arterial wall given advection and diffusion properties. The comparative results of the current study highlight that: (a) among the high and low-dose Everolimus Eluting Balloons, diffusivity is the governing property and as such, respective amount of drug is eluted and bound by the arterial wall - effectively rendering the high-dose balloon powerful in its delivery properties, and (b) based on the diffusivity of Everolimus and Sirolimus, the Everolimus Eluting Balloons exhibit superior performance with respect to total bound drug of the arterial *** Relevance - The approach described in this article is a step farther toward precision drug release from DCB prediction. Furthermore, our study's findings demonstrate that higher drug concentration in the coating leads to higher binding amounts;this information could be utilized to inform the creation of novel Drug Coated Balloon (DCB). © 2024 IEEE.

关键词： Balloons

来源：评论

学校读者我要写书评

暂无评论

HaLViT: Half of the Weights are Enough

HaLViT: Half of the Weights are Enough

引用

IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)

作者： Onur Can Koyun Behçet Uğur Töreyin Dept. of Artificial Intelligence and Data Engineering Informatics Institute Signal Processing for Computational Intelligence Research Group (SP4CING) İstanbul Technical University İstanbul Türkiye

ISBN: (数字)9798350365474

ISBN: (纸本)9798350365481

Deep learning architectures like Transformers and Convolutional Neural Networks (CNNs) have led to ground-breaking advances across numerous fields. However, their extensive need for parameters poses challenges for implementation in environments with limited resources. In our research, we propose a strategy that focuses on the utilization of the column and row spaces of weight matrices, significantly reducing the number of required model parameters without substantially affecting performance. This technique is applied to both Bottleneck and Attention layers, achieving a notable reduction in parameters with minimal impact on model efficacy. Our proposed model, HaLViT, exemplifies a parameter-efficient Vision Transformer. Through rigorous experiments on the ImageNet dataset and COCO dataset, HaLViT’s performance validates the effectiveness of our method, offering results comparable to those of conventional models.

关键词： Computer vision Conferences computational modeling Deep architecture Transformers Pattern recognition Convolutional neural networks

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

Explainable Adversarial Attacks on Coarse-to-Fine Classifier...

引用

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

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

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

关键词： Perturbation methods Decision making Signal processing Acoustics Speech processing

来源：评论

学校读者我要写书评

暂无评论

A Deep Learning Application to Detect and Diagnose Metabolic Encephalopathy

A Deep Learning Application to Detect and Diagnose Metabolic...

引用

Smart Technologies, Communication and Robotics (STCR)

作者： Anjali R Ganesh Babu C Nagarajan N Dept. of Artificial Intelligence and Data Science Coimbatore Institute of Engineering and Technology (Aff.to Anna University ) Coimbatore India Dept. of Electronics and Instrumentation Engineering Bannari Amman Institute of Technology Sathyamangalam India Dept. Electronics and Communication Engineering Coimbatore Institute of Engineering and Technology (Aff.to Anna University ) Coimbatore India

ISBN: (数字)9798350357530

ISBN: (纸本)9798350357547

Brain health plays a major role in human life to recognize their complete prospective over the life course stating the brain functionality which includes sensory, behavioral, cognitive, social-emotional and motor domains regardless of the existence or nonexistence of disorders. If there exists a brain dysfunction due to chemical imbalances in the blood, it is named as Metabolic Encephalopathy. This chemical imbalances occur due to the systematic illness such as diabetes, kidney or liver disease which may disturb mental states and if not treated in right time may lead to permanent brain damage. The role of artificial intelligence in medical field is showing its significant growth in recent years. This document showcases deep learning techniques in the diagnosis of metabolic encephalopathy.

关键词： Training Deep learning Accuracy Systematics Sensitivity Transfer learning Artificial neural networks Artificial intelligence Chemicals Testing

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：