检索结果-内蒙古大学图书馆

34th ACM Web Conference, WWW 2025

作者： Wang, Shuyao Zheng, Zhi Sui, Yongduo Xiong, Hui School of Data Science University of Science and Technology of China Hefei China State Key Laboratory of Cognitive Intelligence University of Science and Technology of China Hefei China Tencent Shenzhen China Thrust of Artificial Intelligence The Hong Kong University of Science and Technology (Guangzhou) Hong Kong Department of Computer Science and Engineering The Hong Kong University of Science and Technology Hong Kong

ISBN: (纸本)9798400712746

Recommender systems are crucial for personalizing user experiences but often depend on implicit feedback data, which can be noisy and misleading. Existing denoising studies involve incorporating auxiliary information or learning strategies from interaction data. However, they struggle with the inherent limitations of external knowledge and interaction data, as well as the non-universality of certain predefined assumptions, hindering accurate noise identification. Recently, large language models (LLMs) have gained attention for their extensive world knowledge and reasoning abilities, yet their potential in enhancing denoising in recommendations remains underexplored. In this paper, we introduce LLaRD, a framework leveraging LLMs to improve denoising in recommender systems, thereby boosting overall recommendation performance. Specifically, LLaRD generates denoising-related knowledge by first enriching semantic insights from observational data via LLMs and inferring user-item preference knowledge. It then employs a novel Chain-of-Thought (CoT) technique over user-item interaction graphs to reveal relation knowledge for denoising. Finally, it applies the Information Bottleneck (IB) principle to align LLM-generated denoising knowledge with recommendation targets, filtering out noise and irrelevant LLM knowledge. Empirical results demonstrate LLaRD’s effectiveness in enhancing denoising and recommendation accuracy. © 2025 Copyright held by the owner/author(s).

关键词： Denoising Large Language Models Recommendation

来源：评论

学校读者我要写书评

暂无评论

Zero-shot Swift Lung 4D CT Segmentation via Optimized Segmentation Anything Model

Zero-shot Swift Lung 4D CT Segmentation via Optimized Segmen...

引用

Medical Imaging 2025: Image Processing

作者： Hu, Mingzhe Qiu, Richard Safari, Mojtaba Roper, Justin Yang, Xiaofeng Department of Radiation Oncology Winship Cancer Institute Emory University AtlantaGA30322 United States Department of Computer Science and Informatics Emory University AtlantaGA30322 United States

ISBN: (纸本)9781510685901

Segmenting lung CT images is crucial for radiation oncology, impacting treatment planning, tumor tracking, and response assessment. This study presents an optimized version of the Segmentation Anything Model (SAM), known as RepViT-SAM, tailored for efficient and accurate lung CT segmentation. RepViT-SAM incorporates the RepViT encoder, which enhances computational efficiency while maintaining high segmentation performance. The model utilizes minimal user input through point prompts to perform zero-shot segmentation, adapting to real-time clinical requirements. We evaluated RepViT-SAM on a 4D lung CT dataset, achieving an average Dice Similarity Coefficient (DSC) of 87.9% and a pixel accuracy of 96.8%, with a significantly reduced latency of 32.7 ms-outpacing both the original SAM and MobileSAM models. These results demonstrate that RepViT-SAM is not only highly accurate but also fast enough to meet the demands of clinical applications, including real-time lung CT segmentation. Future work will focus on expanding the model's capabilities to 3D segmentation and improving its robustness across varying lung morphologies. © 2025 SPIE.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

Speech Separation for Low-Resource Languages

Speech Separation for Low-Resource Languages

引用

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Borsdorf, Marvin Pan, Zexu Himmelmann, Pascal Li, Haizhou Schultz, Tanja Machine Listening Lab University of Bremen Bremen Germany Alibaba Group Singapore Department of Mathematics and Computer Science FernUniversität in Hagen Hagen Germany School of Data Science Shenzhen Research Institute of Big Data The Chinese University of Hong Kong Shenzhen China Department of Electrical and Computer Engineering National University of Singapore Singapore Cognitive Systems Lab University of Bremen Bremen Germany

ISBN: (纸本)9798350368741

Speech separation aims to equip machines with the human ability of selective listening, i.e. to focus attention on specific information in spoken communication. Studies have shown that the language spoken in a cocktail party scenario matters. While the development of speech separation models can leverage extensive databases, for the majority of languages only very limited data is available. This work presents the very first study on speech separation for low-resource languages. We choose blind source separation as the task to be studied and analyze three strategies to overcome the data scarcity of two low-resource languages from the GlobalPhoneMS2 database. We show that data from other languages can be used to develop models that work for low-resource languages. Finetuning additionally boosts the performance, and training on multiple languages increases both performance and robustness. We show that dynamic mixing in the development helps to find a trade-off between performance and development time. © 2025 IEEE.

关键词： cocktail party problem GlobalPhone low-resource language multilingual speech separation

来源：评论

学校读者我要写书评

暂无评论

Q-Rung Orthopair Fuzzy Sets-Enhanced FMEA for COVID-19 Risk Assessment

Informatica (Slovenia)

引用

Informatica (Slovenia) 2025年第1期49卷 81-92页

作者： Abdullah, Lazim Awang, Noor Azzah Qiyas, Muhammad Special Interest Group of Modelling and Data Analytics Faculty of Computer Science and Mathematics Universiti Malaysia Terengganu Kuala Nerus21030 Malaysia College of Computing Informatics and Mathematics Universiti Teknologi MARA Selangor Shah Alam40450 Malaysia Department of Mathematics Riphah International University Failsabad Campus Pakistan

Failure Modes and Effects Analysis (FMEA) is a widely used tool for risk analysis, primarily to identify risk factors affecting system quality. Due to the limitations of the traditional FMEA model, several recent models incorporating advanced fuzzy set extensions have been developed to enhance the reliability of risk assessment outcomes. However, most of these models limit expert flexibility in expressing preferences and often overlook the impact of unequal expert weights and the stability of risk ranking results. This study introduces a new FMEA model based on Q-Rung Orthopair Fuzzy Sets (Q-ROFSs), termed Q-ROFSs-FMEA. Q-ROFSs, an extension of intuitionistic fuzzy sets, introduce a new linguistic term. The Q-ROFSs-FMEA model considers the unequal weights of experts, enabling a dynamic representation of expert preferences. These weights and the linguistic evaluation of risk factors are integrated through an aggregation operator, facilitating consensus among experts. The model is applied to a case study on COVID-19 risk factors, revealing that ‘older age’ (risk priority number 0.000012) is the highest risk factor, while ‘gender’ (risk priority number -0.0037) is the lowest. It is found that the ranking of risk factors determined by the Q-ROFSs-FMEA model is obtained as FM1 FM3 FM6 FM4 FM5 FM7 FM8 FM2 . Furthermore, a comparative analysis indicates consistent ranking results across different models, demonstrating the reliability of the proposed model. The case study and comparative analysis validate the effectiveness and applicability of the Q-ROFSsbased risk assessment model. © 2025 Slovene Society Informatika. All rights reserved.

关键词： Failure modes

来源：评论

学校读者我要写书评

暂无评论

ISWC 2024 Special Session on Harmonising Generative AI and Semantic Web Technologies

引用

CEUR Workshop Proceedings 2025年 3953卷

作者： Alharbi, Reham Berardinis, Jacopo de Groth, Paul Meroño-Peñuela, Albert Simperl, Elena Tamma, Valentina Department of Computer Science University of Liverpool United Kingdom Informatics Institute University of Amsterdam Netherlands Department of Informatics King’s College London United Kingdom

来源：评论

学校读者我要写书评

暂无评论

PRIVACY-PRESERVING GENERATIVE MODELS: A COMPREHENSIVE SURVEY

arXiv

引用

arXiv 2025年

作者： Padariya, Debalina Wagner, Isabel Taherkhani, Aboozar Boiten, Eerke School of Computer Science and Informatics De Montfort University Leicester United Kingdom Department of Mathematics and Computer Science University of Basel Basel Switzerland

Despite the generative model’s groundbreaking success, the need to study its implications for privacy and utility becomes more urgent. Although many studies have demonstrated the privacy threats brought by GANs, no existing survey has systematically categorized the privacy and utility perspectives of GANs and VAEs. In this article, we comprehensively study privacy-preserving generative models, articulating the novel taxonomies for both privacy and utility metrics by analyzing 100 research publications. Finally, we discuss the current challenges and future research directions that help new researchers gain insight into the underlying concepts. Copyright © 2025, The Authors. All rights reserved.

关键词： Differential privacy

来源：评论

学校读者我要写书评

暂无评论

Robust WCET Estimation via Ensemble Learning with Fractional-Order Legendre Function Kernels

Robust WCET Estimation via Ensemble Learning with Fractional...

引用

Applied Research in Electrical Engineering (AREE), National and International Conference on

作者： Amir Hussein Ataei Monfared Mahdi Movahedian Moghaddam Sheedeh Sharif-Bakhtiar Kourosh Parand Keivan Borna Department of Computer Science Faculty of Mathematics and Computer Science Kharazmi University Tehran Iran Department of Computer and Data Sciences Faculty of Mathematical Sciences Shahid Beheshti University Tehran Iran Department of Cognitive Modeling Institute for Cognitive and Brain Sciences Shahid Beheshti University Tehran Iran

ISBN: (数字)9798350377170

ISBN: (纸本)9798350377187

Estimating Worst-Case Execution Time (WCET) as a regression problem has become increasingly challenging due to the complexity of modern hardware and software systems. Traditional statistical methods often fall short of providing accurate and reliable estimates. To address these limitations, ensemble learning methods have emerged as a promising approach for capturing edge cases in execution scenarios. This paper proposes a novel ensemble model that integrates fractional-order Legendre functions (FLF) to enrich feature representation. This integration reduces correlations between individual models within the ensemble, enhancing generalization capabilities. Our method achieves significant reductions in Mean Squared Error (MSE) for WCET estimates while maintaining high levels of safeness. Specifically, the proposed model maintained an error below one across most benchmarks, with safeness levels of $\% 100$ in 13 out of 16 benchmarks. These findings underscore the efficacy of our approach and its potential to provide robust and accurate WCET estimates for complex systems.

关键词： Support vector machines Accuracy Upper bound Estimation Benchmark testing Real-time systems Complexity theory Ensemble learning Time factors Kernel

来源：评论

学校读者我要写书评

暂无评论

Virtual Tool Embodiment in Simulated Gravity Conditions 17

Virtual Tool Embodiment in Simulated Gravity Conditions

引用

17th International Workshop on IMmersive Mixed and Virtual Environment Systems, MMVE 2025, co-located with ACM Multimedia Systems Conference, MMSys 2025

作者： Najafabadi, Amir Jahanian Botev, Jean Sun, Ningyuan Kroger, Carolyn Department of Cognitive Neuroscience Bielefeld University Bielefeld Germany VR/AR Lab Department of Computer Science University of Luxembourg Esch-sur-Alzette Luxembourg Kresge Hearing Research Institute Department of Otolaryngology - Head and Neck Surgery University of Michigan Ann Arbor United States

ISBN: (纸本)9798400714689

This study investigates the impact of different gravitational physics on virtual tool embodiment and examines whether cognitive load predicts the emergence of ownership and agency. We hypothesize that low-gravity environments enhance the sense of tool embodiment and that this enhancement correlates with reduced cognitive load. To this end, we devised a virtual reality experiment in which users both actively and passively experience normal, hyper-, and microgravity conditions during tool use. While the current sample size limits the statistical validation of our hypothesis, preliminary findings indicate promising trends supporting these predictions. These results contribute to a deeper understanding of how altered physical conditions influence virtual interactions. This has potential implications for virtual reality design, human-computer interaction, and space exploration applications. © 2025 Copyright held by the owner/author(s). Publication rights licensed to ACM.

关键词： Virtual environments

来源：评论

学校读者我要写书评

暂无评论

Multilingual Hate Speech Detection: Innovations in Optimized Deep Learning for English and Arabic Hate Speech Detection

引用

SN computer science 2025年第3期6卷 1-16页

作者： AL-Sukhani, Hassan Bsoul, Qusay Elhawary, Abdelrahman H. Nasr, Ziad M. Mansour, Ahmed E. Batyha, Radwan M. Alqadi, Basma S. Alqurni, Jehad Saad Alfagham, Hayat Madbouly, Magda M. Cybersecurity Department Faculty of Science and Information Technology Jadara University Irbid Jordan Cybersecurity Department College of Computer Sciences and Informatics Amman Arab University Amman Jordan Faculty of Engineering Ain Shams University Cairo Egypt Faculty of Computer Science Misr International University Cairo Egypt Faculty of Information Technology Applied Science Private University Amman 11931 Jordan Computer Science Department College of Computer and Information Sciences Al Imam Mohammad Ibn Saud Islamic University Riyadh Saudi Arabia Department of Educational Technologies College of Education Imam Abdulrahman Bin Faisal University P.O. Box 1982 Dammam Saudi Arabia Department of MIS College of Applied Studies and Community Service Imam Abdulrahman Bin Faisal University P.O. Box 1982 Dammam Saudi Arabia Institute of Graduate Studies & Research Alexandria University Alexandria Egypt Faculty of Computer and Data Science Alexandria University Alexandria Egypt

This paper presents the development of a multilingual hate speech detection model that effectively processes and classifies content in both Arabic and English. The study leverages both traditional machine learning models, such as K-Nearest Neighbors (KNN), Naive Bayes, and Support Vector Machines (SVM), as well as advanced deep learning models, specifically Bi-directional Long Short-Term Memory (Bi-LSTM) networks. A key challenge addressed is the classification of mixed-language content, which is common on social media platforms in the MENA region. To enhance detection performance, preprocessing techniques were applied to the text data, and the Synthetic Minority Over-sampling Technique (SMOTE) was used to balance the dataset. The results show that the Bi-LSTM model outperformed traditional machine learning approaches, particularly in identifying hate speech across multiple languages. The proposed model demonstrates superior accuracy and robustness in handling mixed-language input, providing a more effective solution for real-world hate speech detection tasks. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2025.

关键词： Class imbalance Deep learning Hate speech detection Machine learning Mixed Arabic and English training Multilingual NLP Natural language processing (NLP) Sentiment analysis

来源：评论

学校读者我要写书评

暂无评论

Optimizing Heart-Kidney Interaction for Cancer Detection Through Physiological Process Simulation as a Decision Support System

引用

IEEE Access 2025年 13卷 79169-79179页

作者： Jaddi, Najmeh Sadat Abdullah, Salwani Othman, Zalinda Yahya, Yazrina Zakree Ahmad Nazri, Mohd Leng Goh, Say Alvankarian, Fatemeh Alvankarian, Jafar Iranian eUniversity Faculty of Computer Engineering Tehran96351-14578 Iran Universiti Kebangsaan Malaysia Center for Artificial Intelligence Technology Faculty of Information Science and Technology Bangi43600 Malaysia Universiti Kebangsaan Malaysia Center for Software Technology and Management Faculty of Information Science and Technology Bangi43600 Malaysia Universiti Malaysia Sabah Kampus Antarabangsa Labuan Optimisation and Analytics Research Group Faculty of Computing and Informatics Labuan87000 Malaysia University of Maryland Department of Electrical and Computer Engineering College ParkMD20742 United States RayanMiniFab Tehran University Science and Technology Park Tehran*** Iran

A decision support system (DSS) is a computer-based tool used to improve decision-making capabilities for any organization by analyzing the available data. The heart-kidney (HK) model proposed in this paper, as a DSS, simulates the physiological processes of the heart and kidneys in the human body to support cancer detection. The first part of the HK algorithm is inspired by the functions of the heart, where blood molecules (solutions) are sent to the heart, simulating the processes of oxygenation and the pumping of oxygenated blood to the kidneys as a local search. The solutions generated by the heart are then fed into the kidney-inspired algorithm (KA), which performs filtration, reabsorption, secretion, and excretion. The heart component enhances the KA’s local search capability, guiding the algorithm toward optimal solutions. The HK model is evaluated on eight test functions, with results compared to those of the heart algorithm and KA alone. Additionally, the HK algorithm’s effectiveness as a decision support system is tested on benchmark classification and time series prediction problems, with statistical analysis against existing methods in the literature. Finally, the HK algorithm is applied to cancer detection from microRNA data, demonstrating its potential in optimizing decision-making processes in healthcare. © 2013 IEEE.

关键词： Diseases

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：