检索结果-内蒙古大学图书馆

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

作者： Cheng Chu Aishwarya Hastak Fan Chen Luddy School of Informatics Computing and Engineering Indiana University Bloomington USA

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Current quantum generative adversarial networks (QGANs) still struggle with practical-sized data. First, many QGANs use principal component analysis (PCA) for dimension reduction, which, as our studies reveal, can diminish the QGAN’s effectiveness. Second, methods that segment inputs into smaller patches processed by multiple generators face scalability issues. In this work, we propose LSTM-QGAN, a QGAN architecture that eliminates PCA preprocessing and integrates quantum long short-term memory (QLSTM) to ensure scalable performance. Our experiments show that LSTM-QGAN significantly enhances both performance and scalability over state-of-the-art QGAN models, with visual data improvements, reduced Fréchet Inception Distance scores, and reductions of 5× in qubit counts, 5× in single-qubit gates, and 12× in two-qubit gates.

关键词： Visualization Scalability Qubit Logic gates Signal processing Generative adversarial networks Generators Speech processing Long short term memory Principal component analysis

来源：评论

学校读者我要写书评

暂无评论

Bridging Fashion and Technology: Synthetic Human Models for an Enhanced E-Commerce Experience 23rd

Bridging Fashion and Technology: Synthetic Human Models for...

引用

23rd EPIA Conference on Artificial Intelligence, EPIA 2024

作者： Duarte, Ana Rita Conceição, Luís Department of Informatics Engineering School of Engineering Polytechnic of Porto Porto Portugal Research Group on Intelligent Engineering and Computing for Advanced Innovation and Development School of Engineering Polytechnic of Porto Porto Portugal

ISBN: (纸本)9783031734960

The fashion industry is undergoing a significant transformation, driven by advancements in digitalization and artificial intelligence (AI). This paper explores the integration of Stable Diffusion Models (SDMs) and AI to create high-quality images of synthetic models wearing a target cloth in multiple poses, enhancing the users’ experience on e-commerce websites and addressing the fast-paced demands of fashion trends. The proposed pipeline includes multiple steps: face generation, pose estimation, cloth warping, human synthesis, and refinement networks, each designed to enhance the realism and quality of the final images for e-commerce platforms. Experimental results on a small VITON-HD dataset demonstrate this approach’s overall success. BRISQUE, NIQE, and entropy were used for objective evaluation, scoring 14.1449, 4.1354, and 7.1238 respectively, indicating a high level of detail, naturalness, and complexity on the generated images. Future work should focus on enhancing the representation of clothing with complex patterns and lower and full garments. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Clothes

来源：评论

学校读者我要写书评

暂无评论

Can Quality Survive Scale? Toward an Equal-Quality Instance-Dependent Label Noise Model

Can Quality Survive Scale? Toward an Equal-Quality Instance-...

引用

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

作者： Song, Haohao Xiang, Qiao Shu, Jiwu Institute of Artificial Intelligence Xiamen University China Xiamen Key Laboratory of Intelligent Storage and Computing School of Informatics Xiamen University China Minjiang University China

ISBN: (纸本)9798350368741

Label noise model is a technique to construct controlled noisy datasets for evaluating noise-robust algorithms. However, the quality of the generated noise has not been evaluated thoroughly. In this paper, we propose a novel research question: Do the constructed datasets with the same noise rate have equal effects? We answer this question through a carefully designed experiment: We sequentially generate a series of noisy datasets with equal noise rate by excluding the previous noisy samples while controlling the clean samples. Models trained on these datasets show discrepancy generalization performance, indicating the inequality of noise. Our in-depth analysis reveals that the reasons come from (1) the introduction of non-hard samples and (2) the inequality between hard samples. We propose a primary equal-quality instance-dependent label noise model termed EQ-IDN, which alleviates both issues based on the identification of hard samples and stratified sampling. We compare the noise quality generated by EQ-IDN and other models. Experimental results demonstrate that the noise generated by EQ-IDN has lower difference and more stable quality. © 2025 IEEE.

关键词： label noise quality

来源：评论

学校读者我要写书评

暂无评论

Deep-Source EEG Generation with Deep Learning from Different Configurations of Surface EEG

Deep-Source EEG Generation with Deep Learning from Different...

引用

IEEE International Conference on Consumer Electronics (ICCE)

作者： Jiadao Zou Qingxue Zhang Luddy School of Informatics Computing and Engineering Indiana University Department of Computer Science and Department of Biomedical Engineering Indiana University School of Informatics Computing and Engineering Indiana University Luddy Indianapolis

ISBN: (数字)9798331521165

ISBN: (纸本)9798331521172

Brain dynamics big data is of increasing promise for many applications like epilepsy detection and cognitive understanding, with the advancements of consumer technology. However, the deep-source brain measurement is difficult. In this study, we target the brain electroencephalogram (EEG) application, and investigate the deep-source EEG generation from surface EEG towards convenient big data. The deep learning algorithm has been developed to mine different configurations of the surface EEG streams, including the single-channel and multi-channel cases, for deep-source EEG generation. Promising experiments on the epilepsy application have been conducted, demonstrating the great promise of deep-learning-empowered deep-source EEG generation. This study will greatly advance brain dynamics mining towards smart consumer technology.

关键词： Deep learning Heuristic algorithms Smart healthcare Epilepsy Big Data Electroencephalography Streams Consumer electronics

来源：评论

学校读者我要写书评

暂无评论

Towards Robust Federated Learning: Investigating Poisoning Attacks Under Clients Data Heterogeneity 19

Towards Robust Federated Learning: Investigating Poisoning A...

引用

19th International Conference on Ubiquitous Information Management and Communication, IMCOM 2025

作者： Soubih, Abdenour Lahmer, Seyyid Ahmed Abuhamad, Mohammed Abuhmed, Tamer College of Computing and Informatics Sungkyunkwan University Suwon Korea Republic of Department of Information Engineering University of Padova Padua Italy Department of Computer Science Loyola University Chicago Chicago United States

ISBN: (纸本)9798331507817

Federated Learning (FL) offers a privacy-preserving solution by enabling multiple clients to train a shared model collaboratively without centralizing data. However, the decentralized nature of FL presents challenges, particularly regarding security and performance under adversarial conditions. This paper investigates the effects of poisoning attacks under data heterogeneity. Our experiments evaluate the impact of varying malicious client fractions and poison concentration levels on the accuracy of the model. We explore the effects of poisoning attacks on FedAvg and FedNova models using medical imaging tasks. Our findings reveal that increasing data heterogeneity exacerbates the effects of poisoning, with FedNova demonstrating greater resilience compared to FedAvg. We found that the number of malicious clients plays a more significant role in degrading performance than the ratio of poisoning samples shared by each malicious client, suggesting that even modest levels of poisoning can be tolerated by most algorithms. The study highlights the importance of developing robust defense mechanisms to maintain model performance under adversarial conditions. © 2025 IEEE.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Roll for Robot: A Tabletop Role-Playing Game for Designing Socially Assistive Robots for Depression Management 16th

Roll for Robot: A Tabletop Role-Playing Game for Designing ...

引用

16th International Conference on Social Robotics, ICSR + AI 2024

作者： Collins, Sawyer Henkel, Kenna Baugus Henkel, Zachary Śabanović, Selma Luddy School of Informatics Computing and Engineering Indiana University Bloomington BloomingtonIN United States College of Computer Science and Engineering Mississippi State University Mississippi State StarkvilleMS United States

ISBN: (纸本)9789819635214

Imagining and creating a future with a robot tailored to an individual’s needs and wants provides a unique challenge to both designers and perspective users. Using both an existing base zoomorphic robot, TherabotTM in a specially designed tabletop role-playing game, we have explored what it would look like for someone living with depression to go through a "day in the life" with their newly designed robot companion. Inspired by this collaborative co-design technique, eight participants explained their own relationship with depression, identified a character that they connected with, and chose the physical form, sensors, and abilities for a socially assistive robot. Using this character and robot, participants described how they might overcome depression symptoms and identified uses when managing depression symptoms for the robot. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Machine design

来源：评论

学校读者我要写书评

暂无评论

Holistic Cyber Threat Modeling for Machine Learning-Based Systems: A Case Study in Healthcare

Holistic Cyber Threat Modeling for Machine Learning-Based Sy...

引用

18th International Joint Conference on Biomedical Engineering Systems and Technologies, BIOSTEC 2025

作者： Jaal, Janno Bahsi, Hayretdin Department of Software Science Tallinn University of Technology Tallinn Estonia Cybernetica AS Tallinn Estonia School of Informatics Computing and Cyber Systems Northern Arizona University United States

Considering the immense pace in machine learning (ML) technology and related products, it may be difficult to imagine a software system, including healthcare systems, without any subsystem containing an ML model in the near future. However, ensuring the resiliency of these ML-based systems against cyber attacks is vital for more seamless and widespread technology usage. The secure-by-design principle, considering security from the early stages of development, is a cornerstone to achieving sufficient security at a reasonable cost. The realization of this principle starts with conducting threat modeling to understand the relevant security posture and identify cyber security requirements before system design. Although threat modeling of software systems is widely known, it is unclear how to apply it to software systems with machine learning models. Although adversarial machine learning is a widely studied research topic, it has yet to be thoroughly researched how adversarial and conventional cybersecurity attacks can be holistically considered to identify applicable cyber threats at the early stage of a software development life cycle. This paper adapts STRIDE, a widely-known threat modeling method, for the holistic cyber threat analysis of an ML-based healthcare system. © 2025 by SCITEPRESS – Science and Technology Publications, Lda.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Shape-Based Single Object Classification Using Ensemble Method Classifiers

arXiv

引用

arXiv 2025年

作者： Kamarudin, Nur Shazwani Makhtar, Mokhairi Shamsuddin, Syadiah Nor Wan Fadzli, Syed Abdullah Faculty of Informatics and Computing Universiti Sultan Zainal Abidin Tembila Terengganu Malaysia

Nowadays, more and more images are available. Annotation and retrieval of the images pose classification problems, where each class is defined as the group of database images labelled with a common semantic label. Various systems have been proposed for content-based retrieval, as well as for image classification and indexing. In this paper, a hierarchical classification framework has been proposed for bridging the semantic gap effectively and achieving multi-category image classification. A well-known pre-processing and post-processing method was used and applied to three problems;image segmentation, object identification and image classification. The method was applied to classify single object images from Amazon and Google datasets. The classification was tested for four different classifiers;BayesNetwork (BN), Random Forest (RF), Bagging and Vote. The estimated classification accuracies ranged from 20% to 99% (using 10-fold cross validation). The Bagging classifier presents the best performance, followed by the Random Forest classifier. Copyright © 2025, The Authors. All rights reserved.

关键词： Image annotation

来源：评论

学校读者我要写书评

暂无评论

Enhancing Classification Algorithm Recommendation in Automated Machine Learning: A Meta-Learning Approach Using Multivariate Sparse Group Lasso

引用

Computer Modeling in Engineering & Sciences 2025年第2期142卷 1611-1636页

作者： Irfan Khan Xianchao Zhang Ramesh Kumar Ayyasamy Saadat M.Alhashmi Azizur Rahim School of Software Dalian University of TechnologyDalian116620China Faculty of Information and Communication Technology Universiti Tunku Abdul RahmanKampar31900Malaysia College of Computing and Informatics University of SharjahSharjah27272United Arab Emirates Department of Computer Systems Engineering University of Engineering and Applied SciencesSwat19200Pakistan

The rapid growth of machine learning(ML)across fields has intensified the challenge of selecting the right algorithm for specific tasks,known as the Algorithm Selection Problem(ASP).Traditional trial-and-error methods have become impractical due to their resource *** Machine Learning(AutoML)systems automate this process,but often neglect the group structures and sparsity in meta-features,leading to inefficiencies in algorithm recommendations for classification *** paper proposes a meta-learning approach using Multivariate Sparse Group Lasso(MSGL)to address these *** method models both within-group and across-group sparsity among meta-features to manage high-dimensional data and reduce multicollinearity across eight meta-feature *** Fast Iterative Shrinkage-Thresholding Algorithm(FISTA)with adaptive restart efficiently solves the non-smooth optimization *** validation on 145 classification datasets with 17 classification algorithms shows that our meta-learning method outperforms four state-of-the-art approaches,achieving 77.18%classification accuracy,86.07%recommendation accuracy and 88.83%normalized discounted cumulative gain.

关键词： Meta-learning machine learning automated machine learning classification meta-features

来源：评论

学校读者我要写书评

暂无评论

Models for Network Traffic Behavioral Analysis in IoT Systems

Models for Network Traffic Behavioral Analysis in IoT System...

引用

23rd International Conference on informatics in Economy, IE 2024

作者： Zamfiroiu, Alin Floroiu, Iustin Savu, Daniel Bătăgan, Lorena Ghai, Anupriya Sharma National Institute for Research & Development in Informatics Bucharest Romania Bucharest University of Economic Studies Bucharest Romania School of Computing Graphic Era Hill University Dehradun India

ISBN: (纸本)9789819601608

In today's interconnected world, the proliferation of Internet of Things (IoT) devices has revolutionized the way we interact with technology. From smart homes and wearable devices to industrial sensors and autonomous vehicles, IoT systems have permeated nearly every aspect of our daily lives. However, with this exponential growth in IoT adoption comes a host of challenges, particularly concerning the security and performance of these interconnected networks. One of the key methodologies employed to address these challenges is network traffic behavioral analysis. Moreover, network traffic behavioral analysis plays a crucial role in threat intelligence and incident response. By aggregating and analyzing network data over time, organizations can identify recurring attack patterns and indicators of compromise. This rich contextual information enhances their ability to anticipate and preempt future cyber threats. Additionally, in the event of a security incident, network traffic analysis provides invaluable forensic evidence, allowing investigators to reconstruct the sequence of events, attribute responsibility, and mitigate the impact of the breach. In this paper, we propose some models for the network traffic generated by sensors and actuators in IoT systems. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Computer forensics

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：