检索结果-内蒙古大学图书馆

Steel Surface Defect Detection Using Learnable Memory Vision Transformer

Computers, Materials & Continua 2025年第1期82卷 499-520页

作者： Syed Tasnimul Karim Ayon Farhan Md.Siraj Jia Uddin Department of Computer Science and Engineering BRAC UniversityDhaka1212Bangladesh Department of AI and Big Data Endicott CollegeWoosong UniversityDaejeon34606Republic of Korea

This study investigates the application of Learnable Memory Vision Transformers(LMViT)for detecting metal surface flaws,comparing their performance with traditional CNNs,specifically ResNet18 and ResNet50,as well as other transformer-based models including Token to Token ViT,ViT withoutmemory,and Parallel *** awidely-used steel surface defect dataset,the research applies data augmentation and t-distributed stochastic neighbor embedding(t-SNE)to enhance feature extraction and *** techniques mitigated overfitting,stabilized training,and improved generalization *** LMViT model achieved a test accuracy of 97.22%,significantly outperforming ResNet18(88.89%)and ResNet50(88.90%),aswell as the Token to TokenViT(88.46%),ViT without memory(87.18),and Parallel ViT(91.03%).Furthermore,LMViT exhibited superior training and validation performance,attaining a validation accuracy of 98.2%compared to 91.0%for ResNet 18,96.0%for ResNet50,and 89.12%,87.51%,and 91.21%for Token to Token ViT,ViT without memory,and Parallel ViT,*** findings highlight the LMViT’s ability to capture long-range dependencies in images,an areawhere CNNs struggle due to their reliance on local receptive fields and hierarchical feature *** additional transformer-based models also demonstrate improved performance in capturing complex features over CNNs,with LMViT excelling particularly at detecting subtle and complex defects,which is critical for maintaining product quality and operational efficiency in industrial *** instance,the LMViT model successfully identified fine scratches and minor surface irregularities that CNNs often *** study not only demonstrates LMViT’s potential for real-world defect detection but also underscores the promise of other transformer-based architectures like Token to Token ViT,ViT without memory,and Parallel ViT in industrial scenarios where complex spatial relationships are *** research m

关键词： Learnable Memory Vision Transformer(LMViT) Convolutional Neural Networks(CNN) metal surface defect detection deep learning,computer vision image classification learnable memory gradient clipping label smoothing t-SNE visualization

来源：评论

学校读者我要写书评

暂无评论

Harnessing the Power of Attention for Patch-Based Biomedical Image Classification

引用

IEEE Access 2025年 13卷 36760-36774页

作者： Habib, Gousia Ahmad Malik, Ishfaq Qureshi, Shaima Ra, In-Ho Singh, Saurabh Sharma, Surbhi Department of Electrical Engineering New Delhi110016 India Shoolini University Yogananda School of AI Computer and Data Sciences Solan173229 India Department of Computer Science Engineering Srinagar190006 India Kunsan National University School of Computer Information and Communication Engineering Gunsan54150 Korea Republic of Woosong University Department of AI and Big Data Daejeon34606 Korea Republic of

The analysis of biomedical images, particularly brain MRI scans, is critical in healthcare and medical research. However, conventional approaches such as convolutional neural networks (CNNs) often struggle to capture complex spatial and contextual relationships within medical imaging data, limiting their generalization capabilities. To address these limitations, we propose a novel framework for brain MRI analysis using Vision Transformers (ViTs) augmented with a Shifted Patching Technique (S.P.T.). This approach combines conventional non-overlapping patching with shifted patching to better model intricate spatial dependencies and contextual associations. The proposed method is evaluated on the BRaiN MRI dataset, with standardized image resolution achieved through Lanczos5 interpolation. Further validation is performed on the ChestX-ray14 and Camelyon17 datasets to assess its generalizability across medical imaging modalities. Performance metrics such as accuracy, F1-score, and precision demonstrate that the proposed ViT model consistently outperforms traditional CNN-based methods. Ablation studies further confirm the significant contribution of the S.P.T. and other architectural components to overall performance. These findings highlight the potential of ViT-based architectures with innovative patching techniques as powerful tools for biomedical image analysis. This work offers a robust and scalable solution for enhancing diagnostic accuracy and advancing medical research applications. © 2013 IEEE.

关键词： Diagnosis

来源：评论

学校读者我要写书评

暂无评论

Evaluating ai-powered predictive solutions for MRI in lumbar spinal stenosis: a systematic review

引用

Artificial Intelligence Review 2025年第8期58卷 1-78页

作者： Al-antari, Mugahed A. Salem, Saied Raza, Mukhlis Elbadawy, Ahmed S. Bütün, Ertan Aydin, Ahmet Arif Aydoğan, Murat Ertuğrul, Bilal Talo, Muhammed Gu, Yeong Hyeon Department of Artificial Intelligence and Data Science College of AI Convergence Daeyang AI Center Sejong University Seoul05006 Korea Republic of Department of Artificial Intelligence College of AI Convergence Daeyang AI Center Sejong University Seoul05006 Korea Republic of Department of Computer Engineering Faculty of Engineering Fırat University Elazığ Turkey Department of Computer Engineering Faculty of Engineering Inonu University Malatya Turkey Department of Software Engineering Faculty of Technology Fırat University Elazığ Turkey Department of Neurosurgery Faculty of Medicine Fırat University Elazığ Turkey Department of Computer Science & Engineering University of North Texas DentonTX76205 United States

Lumbar spinal stenosis (LSS) involves the narrowing of the spinal canal, leading to compression of the spinal cord and nerves in the lower back. Common causes include injuries, degenerative age-related changes, congenital conditions, and tumors, all of which contribute to back pain. Early diagnosis is critical for symptom management, preventing progression, and preserving quality of life. This study systematically reviews ai-based approaches for predicting LSS using MRI axial and sagittal imaging. The review focuses on various ai tasks: detection, segmentation, classification, hybrid approaches, spinal index measurements (SIM), and explainable ai frameworks. The aim is to highlight current knowledge, identify limitations in existing models, and propose future research directions. Following PRISMA guidelines and the PICO method (Population, Intervention, Comparison, Outcome), the review collects data from databases like PubMed, Web of Science, ScienceDirect, and IEEE Xplore (2005–2024). The Rayyan ai tool is used for duplicate removal and screening. The screening process includes an initial review of titles and abstracts, followed by full-text appraisal. The Meta Quality Appraisal Tool (MetaQAT) assesses the quality of selected articles. Of 1323 records, 97 duplicates were removed. After screening, 895 records were excluded, leaving 331 for full-text review. Among these, 184 articles were excluded for lacking ai relevance. Ultimately, 95 key articles (91 technical papers and 4 reviews) were identified for their contributions to ai-based LSS prediction. This review provides a comprehensive analysis of ai techniques in LSS prediction, guiding future research and advancing understanding in areas like explainable ai and large language models (LLMs). © The Author(s) 2025.

关键词： Explainable artificial intelligence (Xai) Harmonization Large language models (LLMs) LSS prediction Lumbar spinal stenosis (LSS) Spinal LSS indices measurements

来源：评论

学校读者我要写书评

暂无评论

Model-based framework for exploiting sensors of IoT devices using a botnet: a case study with android

引用

Cyber-Physical Systems 2025年第1期11卷 1-46页

作者： Khaliq, Zubair Khan, Dawood Ashraf Baba, Asif Iqbal Ali, Shahbaz Farooq, Sheikh Umar Department of Computer Science University of Kashmir Jammu and Kashmir Srinagar India AI Thndr Dubai United Arab Emirates Department of Computer Science Tuskegee University TuskegeeAL United States Data Engineering hyke.ai Dubai United Arab Emirates

Botnets have become a severe security threat not only to the Internet but also to the devices connected to it. Factors like the exponential growth of IoT, the COVID-19 pandemic, and the ever-larger number of cybercriminals are incentivising the growth of botnets in this domain. The recent outbreak of botnets like Dark Nexus (derived from Qbot and Mirai), Mukashi, LeetHozer, Hoaxcalls, etc. shows the alarming rate at which this threat is converging. The botnets have attributes that make them an excellent platform for malicious activities in IoT devices used by organisations to safeguard the personal and confidential data of their customers, employees, and business partners. The IoT devices have built-in sensors or actuators that can be exploited to monitor or control the physical environment of the entities connected to them, thereby violating the fundamental concept of privacy-by-design of these devices. In this paper, we design and describe a modular botnet framework for IoT. The framework uses an enhanced centralised architecture associated with a novel ‘Domain Fluxing Technique’. The proposed framework will provide insights into how privacy in IoT devices can be incorporated at design time to check the sensors and actuators in these devices against malicious exploitation, consequently preserving privacy. © 2024 Informa UK Limited, trading as Taylor & Francis Group.

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

A Distributed Neural Network Architecture for Dynamic Sensor Selection With Application to Bandwidth-Constrained Body-Sensor Networks

引用

IEEE Journal of Biomedical and Health Informatics 2025年第5期29卷 3465-3477页

作者： Strypsteen, Thomas Bertrand, Alexander KU Leuven Department of Electrical Engineering STADIUS Center for Dynamical Systems Signal Processing and Data Analytics LeuvenB-3001 Belgium KU Leuven Leuven.AI KU Leuven institute for AI LeuvenB-3001 Belgium

We propose a dynamic sensor selection approach for deep neural networks (DNNs), which is able to derive an optimal sensor subset selection for each specific input sample instead of a fixed selection for the entire dataset. This dynamic selection is jointly learned with the task model in an end-to-end way, using the Gumbel-Softmax trick to allow the discrete decisions to be learned through standard backpropagation. We then show how we can use this dynamic selection to increase the lifetime of a wireless sensor network (WSN) by imposing constraints on how often each node is allowed to transmit. We further improve performance by including a dynamic spatial filter that makes the task-DNN more robust against the fact that it now needs to be able to handle a multitude of possible node subsets. Finally, we explain how the selection of the optimal channels can be distributed across the different nodes in a WSN. We validate this method on a use case in the context of body-sensor networks, where we use real electroencephalography (EEG) sensor data to emulate an EEG sensor network for motor execution decoding. For this use case, we demonstrate that the distributed algorithm -with only a small amount of cooperation between the nodes- achieves a performance close to the upper bound defined by a fully centralized dynamic selection (maximum absolute decrease of 4% in accuracy). Furthermore, we observe that our dynamic sensor selection framework can achieve large reductions in transmission energy with a limited cost to the task accuracy, validating it as a practical tool for increasing the lifetime of body-sensor networks. © 2025 IEEE.

关键词： Sensor nodes

来源：评论

学校读者我要写书评

暂无评论

Solid-SQL: Enhanced Schema-linking based In-context Learning for Robust Text-to-SQL 31

Solid-SQL: Enhanced Schema-linking based In-context Learning...

引用

31st International Conference on Computational Linguistics, COLING 2025

作者： Liu, Geling Tan, Yunzhi Zhong, Ruichao Xie, Yuanzhen Zhao, Lingchen Wang, Qian Hu, Bo Li, Zang School of Cyber Science and Engineering Wuhan University China Big Data and AI Platform Department Tencent China

ISBN: (纸本)9798891761964

Recently, large language models (LLMs) have significantly improved the performance of text-to-SQL systems. Nevertheless, many state-of-the-art (SOTA) approaches have overlooked the critical aspect of system robustness. Our experiments reveal that while LLM-driven methods excel on standard datasets, their accuracy is notably compromised when faced with adversarial perturbations. To address this challenge, we propose a robust text-to-SQL solution, called Solid-SQL, designed to integrate with various LLMs. We focus on the pre-processing stage, training a robust schema-linking model enhanced by LLM-based data augmentation. Additionally, we design a two-round, structural similarity-based example retrieval strategy for in-context learning. Our method achieves SOTA SQL execution accuracy levels of 82.1% and 58.9% on the general Spider and Bird benchmarks, respectively. Furthermore, experimental results show that Solid-SQL delivers an average improvement of 11.6% compared to baselines on the perturbed Spider-Syn, Spider-Realistic, and Dr. Spider benchmarks. © 2025 Association for Computational Linguistics.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

A survey on cross-user federated recommendation

引用

Science China(Information Sciences) 2025年第4期68卷 7-32页

作者： Enyue YANG Yudi XIONG Wei YUAN Weike PAN Qiang YANG Zhong MING College of Computer Science and Software Engineering Shenzhen University School of Electrical Engineering and Computer Science The University of Queensland WeBank AI Lab WeBank Department of Computer Science and Engineering Hong Kong University of Science and Technology College of Big Data and Internet Shenzhen Technology University Guangdong Laboratory of Artificial Intelligence and Digital Economy (SZ)

Recommender systems are effective in mitigating information overload, yet the centralized storage of user data raises significant privacy concerns. Cross-user federated recommendation(CUFR) provides a promising distributed paradigm to address these concerns by enabling privacy-preserving recommendations directly on user devices. In this survey, we review and categorize current progress in CUFR, focusing on four key aspects: privacy, security, accuracy, and efficiency. Firstly,we conduct an in-depth privacy analysis, discuss various cases of privacy leakage, and then review recent methods for privacy protection. Secondly, we analyze security concerns and review recent methods for untargeted and targeted *** untargeted attack methods, we categorize them into data poisoning attack methods and parameter poisoning attack methods. For targeted attack methods, we categorize them into user-based methods and item-based methods. Thirdly,we provide an overview of the federated variants of some representative methods, and then review the recent methods for improving accuracy from two categories: data heterogeneity and high-order information. Fourthly, we review recent methods for improving training efficiency from two categories: client sampling and model compression. Finally, we conclude this survey and explore some potential future research topics in CUFR.

关键词： cross-user federated recommendation federated recommendation federated learning recommender systems user privacy

来源：评论

学校读者我要写书评

暂无评论

Domain Separation aided data Efficiency Enhancement for Thermal Object Detection

Domain Separation Aided Data Efficiency Enhancement for Ther...

引用

2025 International Conference on Electronics, Information, and Communication, ICEIC 2025

作者： Cho, Beomsik Choi, Seoyeong Lee, Soomok Networks and AI Ajou University Department of Data Suwon Korea Republic of Ajou University Department of Mechanical Engineering Suwon Korea Republic of Ajou University Department of Mobility Engineering Suwon Korea Republic of

ISBN: (纸本)9798331510756

Object detection is a critical component in autonomous driving systems, requiring robust performance across diverse lighting conditions, including nighttime scenarios where RGB cameras underperform due to low visibility. Thermal imaging overcomes this limitation by capturing infrared radiation, facilitating reliable object detection in low-light conditions. However, the development of deep learning models for thermal object detection is hindered by the scarcity of large-scale annotated thermal datasets. To tackle this challenge, we propose a novel approach that enhances data efficiency for thermal object detection through domain separation techniques. Our method introduces a Domain Invariant Encoder (DIE) and a Domain Specific Encoder (DSE). The DIE maps thermal images into a shared feature space with RGB images by minimizing the H-divergence between their feature distributions via adversarial learning. This alignment facilitates the transfer of knowledge from pre-trained RGB models to the thermal domain. Concurrently, the DSE captures unique thermal characteristics by learning domain-specific features. We integrate the invariant and specific features using a cross-attention mechanism, allowing the model to leverage shared and unique information effectively. Evaluated on the FLIR ADAS dataset, our approach achieves state-of-the-art performance with limited thermal data, demonstrating significant improvements in mean Average Precision (mAP) over baseline models. This study addresses the challenge of limited annotated data for thermal object detection, advancing thermal imaging applications in autonomous driving. © 2025 IEEE.

关键词： domain specific object detection thermal object detection

来源：评论

学校读者我要写书评

暂无评论

MedEx: Enhancing Medical Question-Answering with First-Order Logic based Reasoning and Knowledge Injection 31

MedEx: Enhancing Medical Question-Answering with First-Order...

引用

31st International Conference on Computational Linguistics, COLING 2025

作者： Zafar, aizan Mishra, Kshitij Ekbal, Asif Department of Computer Science and Engineering Indian Institute of Technology Patna India School of AI and Data Science Indian Institute of Technology Jodhpur India

ISBN: (纸本)9798891761964

In medical question-answering, traditional knowledge triples often fail due to superfluous data and their inability to capture complex relationships between symptoms and treatments across diseases. This limits models' ability to provide accurate, contextually relevant responses. To overcome this, we introduce MedEx, which employs First-Order Logic (FOL)-based reasoning to model intricate relationships between diseases and treatments. We construct FOL-based triplets that encode the interplay of symptoms, diseases, and treatments, capturing not only surface-level data but also the logical constraints of the medical domain. MedEx encodes the discourse (questions and context) using a transformer-based unit, enhancing context comprehension. These encodings are processed by a Knowledge Injection Cell that integrates knowledge graph triples via a Graph Attention Network. The Logic Fusion Cell then combines medical-specific logical rule triples (e.g., co-occurrence, causation, diagnosis) with knowledge triples and extracts answers through a feed-forward layer. Our analysis demonstrates MedEx's effectiveness and generalization across medical question-answering tasks. By merging logical reasoning with knowledge, MedEx provides precise medical answers and adapts its logical rules based on training data nuances. © 2025 Association for Computational Linguistics.

关键词： Question answering

来源：评论

学校读者我要写书评

暂无评论

Extending LLMs to New Languages: A Case Study of Llama and Persian Adaptation 31

Extending LLMs to New Languages: A Case Study of Llama and P...

引用

31st International Conference on Computational Linguistics, COLING 2025

作者： Sani, Samin Mahdizadeh Sadeghi, Pouya Vu, Thuy-Trang Yaghoobzadeh, Yadollah Haffari, Gholamreza Department of Electrical and Computer Engineering University of Tehran Iran Tehran Institute for Advanced Studies Khatam University Iran Department of Data Science & AI Monash University Australia

ISBN: (纸本)9798891761964

Large language models (LLMs) have made great progress in classification and text generation tasks. However, they are mainly trained on English data and often struggle with low-resource languages. In this study, we explore adding a new language, i.e., Persian, to Llama (a model with a limited understanding of Persian) using parameter-efficient fine-tuning. We employ a multi-stage approach involving pretraining on monolingual Persian data, aligning representations through bilingual pretraining and instruction datasets, and instruction-tuning with task-specific datasets. We evaluate the model's performance at each stage on generation and classification tasks. Our findings suggest that incorporating the Persian language, through bilingual data alignment, can enhance classification accuracy for Persian tasks, with no adverse impact and sometimes even improvements on English tasks. Additionally, the results highlight the model's initial strength as a critical factor when working with limited training data, with cross-lingual alignment offering minimal benefits for the low-resource language. Knowledge transfer from English to Persian has a marginal effect, primarily benefiting simple classification tasks. © 2025 Association for Computational Linguistics.

关键词： Computational linguistics

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：