检索结果-内蒙古大学图书馆

${\sf Img2Acoustic}$Img2Acoustic: A Cross-Modal Gesture Recognition Method Based on Few-Shot Learning

IEEE Transactions on Mobile Computing 2024年第3期24卷 1496-1512页

作者： Yongpan Zou Jianhao Weng Wenting Kuang Yang Jiao Victor C. M. Leung Kaishun Wu College of Computer Science and Software Engineering Shenzhen University Shenzhen China Faculty of Computer Science and Control Engineering Shenzhen University of Advanced Technology Shenzhen China Artificial Intelligence Research Institute Shenzhen MSU-BIT University Shenzhen China Information Hub Hong Kong University of Science and Technology Guangzhou China

Acoustic-based human gesture recognition (HGR) offers diverse applications due to the ubiquity of sensors and touch-free interaction. However, existing machine learning approaches require substantial training data, making the process time-consuming, costly, and labor-intensive. Recent studies have explored cross-modal methods to reduce the need for large training datasets in behavior recognition, but they typically rely on open-source datasets that closely align with the target domain, limiting flexibility and complicating data collection. In this paper, we propose ${\sf Img2Acoustic}$ , a novel cross-modal acoustic-based HGR approach that leverages models trained on open-source image datasets (i.e., EMNIST, Omniglot) to effectively recognize custom gestures detected via acoustic signals. Our model incorporates a task-aware attention layer (TAAL) and a task-aware local matching layer (TALML), enabling seamless transfer of knowledge from image datasets to acoustic gesture recognition. We implement ${\sf Img2Acoustic}$ on commercial devices and conduct comprehensive evaluations, demonstrating that our method not only delivers superior accuracy and robustness compared to existing approaches but also eliminates the need for extensive training data collection.

关键词： Training Gesture recognition Acoustics Data models Sensors Feature extraction Target recognition Mobile computing Image recognition Few shot learning

来源：评论

学校读者我要写书评

暂无评论

A Novel Text Adversarial Sample Generation and Defense Method for SIoT Systems

引用

IEEE Internet of Things Journal 2024年

作者： Ji, Hangxu Guo, Jiale Sun, Yongjiao Yuan, Ye Wang, Guoren Wang, Qi Northeastern University College of Information Science and Engineering Shenyang110819 China Northeastern University Foshan Graduate School of Innovation Foshan528311 China Northeastern University School of Computer Science and Engineering Shenyang110819 China Beijing Institute of Technology School of Computer Science and Technology Beijing100081 China

The generation and defense of text adversarial samples are crucial for improving the robustness and security of SIoT systems, as the exchange of information between devices in SIoT relies heavily on NLP technology. However, the discrete nature of text data leads to a lack of contextual integration in current mainstream adversarial sample generation methods based on text replacement. This results in poor stealth of the generated samples and inefficiencies due to excessive queries to the target model. Meanwhile, defense methods like adversarial training are insufficiently universal and generalizable to handle the diverse and complex range of adversarial attack strategies. This paper proposes a contrastive learning-based method for generating text adversarial samples and a mutual information regularization-based method for defending against text adversarial samples, tailored to the characteristics of devices in SIoT systems and the challenges mentioned above. The proposed method leverage keyword localization, optimal perturbation, and candidate set evaluation to enhance the effectiveness of adversarial samples. Additionally, by combining mutual information measures, statistical estimation functions, and idempotent constraints, the model itself is equipped with effective defenses against adversarial samples. Compared to the baseline, the proposed method significantly reduce the magnitude of perturbations and the number of access attempts to the original samples, while greatly increasing the attack success rate. When applying the proposed adversarial sample defense method, the model's accuracy showed a significant improvement after being subjected to adversarial attacks. © 2024 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Influence Analysis of the Screen Time on Daily Exercise Based on the Personal Activity Factor Model 10

Influence Analysis of the Screen Time on Daily Exercise Base...

引用

10th IEEE Region 10 Humanitarian technology Conference, R10-HTC 2022

作者： Sekine, Yusuke Kasuya, Seiji Tago, Kiichi Chiba Institute of Technology Faculty of Information and Computer Science Narashino Japan Advanced Research Center for Human Sciences Waseda University Tokorozawa Japan

ISBN: (数字)9781665401562

ISBN: (纸本)9781665401562

The lockdown caused by the COVID-19 epidemic has led to using smartphones and decreasing physical activity in the world. It is known that increased screen time and decreased physical activity have bad effects on physical and mental health. However, few studies investigate the influence of screen time on the amount of exercise. By analyzing the influence, exercise promotion services can be improved. Therefore, in this study, we challenge to clarify the relationship between exercise and the screen time. Using a machine learning method, we verify the influence importance of screen time on the amount of exercise. We collect the data on smartphone screen time, tablet screen time, steps, and sleep score from a male college student. In addition, we gather weather of the location, and weekdays/weekends data during our experiment period. To analyze the influence, Gradient Boosting Decision Tree (GBDT) is used. GBDT is a kind of decision tree-based method, and it can show the importance of explainable variables. As a result, the variables that were more important for exercise were, in order of importance, total screen time (total screen times of his smartphone and tablet), screen time of smartphone, and sleep score. The result showed that using electronic devices such as smartphones and tablets influence on the exercise. Moreover, the sleep score also had an influence. © 2022 IEEE.

关键词： Smartphones

来源：评论

学校读者我要写书评

暂无评论

Shadow learner system: implementation of CNN with explainable AI model for bone radiology image classification

引用

Soft Computing 2025年第7期29卷 3571-3588页

作者： Wu, Yaoyang Fong, Simon Liu, Liansheng Department of Computer and Information Science University of Macau Taipa China Medical Imaging Department The Sixth Affiliated Hospital of Jinan University Guangdong China Faculty of Digital Science and Technology Macau Millennium College China

Advancements in deep learning models for medical image classification have greatly improved accuracy, yet the challenge of explainability persists, making it difficult for clinicians to trust and effectively utilize these models in decision-making. This paper presents the Shadow Learner System, a novel approach that combines Convolutional Neural Networks with Explainable AI (XAI) techniques, focusing on the value of misclassified data—both false positives and negatives—by analyzing them to uncover important insights that traditional methods might overlook. We call our framework the Duplex computer-Aided Diagnosis System. Using the Musculoskeletal Radiographs dataset from Stanford University, which includes images of bone abnormalities that may require surgical intervention, we demonstrate how our system enhances explainability. Our findings show that about 85% of the Regions of Interest identified from misclassified images provided useful insights that clinicians found valuable in their work. Moreover, we observed that the shadow set—comprising important misclassified instances—constituted approximately 4–6% of the original dataset. This research highlights how misclassified data can reveal hidden patterns that improve the overall understanding and trust in AI models used in healthcare. The Shadow Learner System not only serves as a reference for treatment decisions but also has potential applications for other medical conditions that require surgical evaluation, such as osteonecrosis and various tumors. By shifting the focus of XAI from mere model validation to enhancing clinical interpretability, this work contributes to the more effective integration of AI in medical practice. © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2025.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Mode Collapse Alleviation of Reinforcement Learning-based GANs in Drug Design

Mode Collapse Alleviation of Reinforcement Learning-based GA...

引用

IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

作者： Zongli Jiang Zhenbo Wang Jinli Zhang Man Wub Chen Li Yoshihiro Yamanishi Faculty of Information Technology Beijing University of Technology China Department of Information and Computer Science Keio University Japan Graduate School of Informatics Nagoya University Japan

De novo drug design is a challenging task that involves understanding the principles of chemistry, chemical properties, and the rules that govern molecular interactions. Deep learning-based generative models, such as MolGAN, offer a promising approach for generating new molecules with the desired chemical properties from molecular graphs. Such models often combine a discrete generative adversarial network (GAN) and reinforcement learning (RL) to produce highly valid and novel molecules. However, the severe mode collapse problem leads to low performance. This study aims to alleviate and investigate the effect of multiple factors on mode collapse. We conducted experiments on different sampling methods, training epochs, and datasets of various volumes and evaluated the experimental results using performance metrics such as validity, uniqueness, novelty, and diversity. The experimental results demonstrate that noise sampling distributions, training epochs, and training data volumes affect performance. The experimental results provide a direction for mitigating the mode collapse problem for RL-based discrete GANs.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Exploring Low Rate Dos Attack Detection Methods: A Bibliometric Analysis and Roadmap for Further Research

Exploring Low Rate Dos Attack Detection Methods: A Bibliomet...

引用

2023 IEEE International Conference on Data and Software Engineering, ICoDSE 2023

作者： Widodo, Anteng Rochim, Adian Fatchur Warsito, Budi School of Postgraduated Studies Diponegoro University Doctoral Program of Information System Semarang Indonesia Muria Kudus University Faculty of Engineering Department of Information System Kudus Indonesia Diponegoro University Faculty of Engineering Department of Computer Engineering Semarang Indonesia Diponegoro University Faculty of Science and Mathematics Department of Statistics Semarang Indonesia

ISBN: (纸本)9798350381382

Low-Rate Denial of Service (LDoS) attacks, an emerging breed of DoS attacks, present a formidable challenge in terms of their detectability. Within the realm of network security, these attacks cast a substantial shadow over the availability and integrity of services. This study seeks to delve into the intricate mechanics of LDoS attacks and explore the landscape of developed detection methodologies. Employing bibliometric analysis, we employ a multifaceted approach to gather research insights pertaining to LDoS attack detection. This approach encompasses the examination of evolving research trends, scrutiny of widely cited scientific publications within the same domain, identification of prolific researchers shaping the field, and a judicious selection of reputable publishers for dependable information acquisition. Furthermore, our study undertakes a meticulous categorization and classification of previously proposed LDoS detection methods documented in well-established scientific literature. This endeavor not only provides a more profound comprehension of the evolutionary trajectory and thematic thrusts within LDoS attack detection research, but it also facilitates the identification of gaps and directions for future exploration. The culmination of these analytical efforts culminates in a comprehensive framework. This contribution manifests through structured categorization, classification, and a forward-looking research roadmap, which collectively serve as valuable tools for both current and aspiring researchers engaged in LDoS attack detection. In summary, this research not only delves into the complexities of LDoS attacks and their mechanisms but also augments the current understanding of detection capabilities. Through its nuanced contributions, this paper extends a helpful guide to the dynamic realm of LDoS attack detection, thus catalyzing advancements and insights for the future. copy;2023 IEEE. © 2023 IEEE.

关键词： Denial-of-service attack

来源：评论

学校读者我要写书评

暂无评论

Optimizing Wireless Sensor Network Topology with Deep Reinforcement Learning for Multi-Source/Destination Scenarios

Optimizing Wireless Sensor Network Topology with Deep Reinfo...

引用

IEEE/ACM International Symposium on Big Data Computing (BDC)

作者： Kosei Kobayashi Masayoshi Aritsugi Pedro Henrique González Israel Mendonça Graduate School of Science and Technology Kumamoto University Kumamoto Japan Faculty of Advanced Science and Technology Kumamoto University Kumamoto Japan System Engineering and Computer Science Program Federal University of Rio de Janeiro Rio de Janeiro Brazil

ISBN: (数字)9798350367300

ISBN: (纸本)9798350367317

Wireless sensor networks are widely valued for their effectiveness in real-time data collection. As the amount of data exchanged within such networks grows, designing a robust network topology that maximizes area coverage with minimal sensors has become a critical challenge. The choice of topology impacts key network metrics, including sensor coverage, communication range, connectivity, inference, and installation and management costsIn this paper, we address the Wireless Sensor Network Planning Problem with Multiple Sources/Destinations, presenting an optimization approach based on deep reinforcement learning. This problem is noteworthy, as sensors in various applications are often required to share data within distinct destinationsWe leverage deep reinforcement learning to effectively address the complex task of selecting optimal sensor locations. Our reinforcement learning agent dynamically learns network structure by iteratively adding and removing sensors, optimizing both sensor coverage and the total number of sensors used. Experiment across diverse scenarios demonstrate the effectiveness of our method for network planning problems of varying scales, achieving full coverage with fewer sensors than traditional approaches. Additionally, our approach also produce solutions for large instances where Mixed Integer programming solvers were not able to. Overall, our method was able to reduce the number of sensors used by up to 22.3% compared to other methods.

关键词： Measurement Integer programming Wireless sensor networks Network topology Deep reinforcement learning Real-time systems Sensors Planning Topology Optimization

来源：评论

学校读者我要写书评

暂无评论

Efficient Hyperparameter Optimization Using Deep Q-Network and BRKGA

Efficient Hyperparameter Optimization Using Deep Q-Network a...

引用

Congress on Evolutionary Computation

ISBN: (数字)9798350308365

ISBN: (纸本)9798350308372

Hyperparameter optimization (HPO) is paragon to maximize performance when designing machine learning models. Among different HPO methods, Genetic Algorithm (GA) based optimization is considered effective because it allows a wide and diverse range of solutions to be explored. However, GA's exploratory nature makes this type of algorithm to evaluate many solutions that do not improve the overall performance. This is specially costly when the objective function to be evaluated is time-consuming, like in the HPO field. In this paper, we propose an efficient hybrid algorithm that is able to reduce computational cost by combining deep reinforcement learning with the Biased Random Key Genetic Algorithm (BRKGA), a variant of genetic algorithms. Our reinforcement learning agent has a decision-making role during the population's fitness calculation, in which it filters out chromosomes that would not improve the overall fitness of the population. The agent uses small amounts of pre-trained data to identify trends in potentially good solutions, and carry out its decision process. We conduct experiments on eight different datasets to assess the effectiveness of the proposed method, and the results show that the proposed method can significantly reduce the computation time of hyperparameter search using BRKGA (up to 44% reduction in computational time) without compromising the quality of the solution (no statistically difference in results).

关键词： Machine learning algorithms Filters Sociology Evolutionary computation Hyperparameter optimization Market research Linear programming

来源：评论

学校读者我要写书评

暂无评论

HistoryQuest: Arabic Question Answering in Egyptian History with LLM Fine-Tuning and Transformer Models

HistoryQuest: Arabic Question Answering in Egyptian History ...

引用

Intelligent Methods, Systems, and Applications (IMSA)

作者： Samaa Maged Asmaa ElMaghraby Ali Marzban Mohamed Essawey Amira Ahmed Esraa Negm Wael H. Gomaa Artificial Intelligence Program School of Information Technology and Computer Science (ITCS) Nile University Giza Egypt Faculty of Computers and Artificial Intelligence Beni-Suef University

ISBN: (数字)9798350362633

ISBN: (纸本)9798350362640

Question answering (QA) in Egyptian history presents a unique and complex challenge for Arabic natural language processing (NLP). This study aims to explore and assess how large language models (LLMs) can enhance the accuracy and performance of Arabic question answering (QA), specifically in this domain. To conduct this investigation, we utilize two comprehensive datasets: the Arabic History-QA dataset and the Contextual Articles Dataset, which cover pivotal historical periods. We evaluate transformer-based models, including AraBERTv2, BERT-large-Arabic with Retrieval-Augmented Generation (RAG), fine-tuned LLaMa-2, and zero-shot LLaMa-3 with Retrieval-Augmented Generation (RAG). Through a rigorous and detailed evaluation process, we analyze how these models address various questions related to Egyptian history. This research contributes valuable insights into advancing the capabilities of Arabic NLP in specialized domains such as historical question answering. Our best results, summarized as the superiority of LLMs, beat those with transformers; additionally, the RAG significantly raised the performance level overall.

关键词： Analytical models Accuracy Large language models Transformers Question answering (information retrieval) History

来源：评论

学校读者我要写书评

暂无评论

Optimization of Node Reduction Using BRKGA for GNN-Based Traffic Speed Forecasting

Optimization of Node Reduction Using BRKGA for GNN-Based Tra...

引用

Congress on Evolutionary Computation

作者： Yuto Inokuchi Pedro Henrique González Israel Mendonça Masayoshi Aritsugi Graduate School of Science and Technology Kumamoto University Kumamoto Japan System Engineering and Computer Science Program Federal University of Rio de Janeiro Rio de Janeiro Brazil Faculty of Advanced Science and Technology Kumamoto University Kumamoto Japan

ISBN: (数字)9798350308365

ISBN: (纸本)9798350308372

While Spatio-Temporal Graph Convolutional Networks (STGCNs) are an effective method for traffic speed fore-casting, their training and inference tend to be time-consuming. In this paper, we aim to refine these networks by strategically reducing their number of nodes, thereby boosting computational efficiency. The nodes in these graphs represent data observed for road segments, and by analyzing the interconnections and layout of the graph, we can identify nodes with minimal contribution to overall performance. Removing these nodes can potentially decrease computation time while maintaining the prediction accuracy. We employ the Biased Random-Key Genetic Algorithm (BRKGA) to identify a good set of nodes for removal, based on a predefined percentage reduction of the original graph size (e.g., retaining 95 % of the original graph). We evaluate different graph size configurations, ranging from 95 % to 70 % node retention, to determine the least impactful node set performance. Our experiments on three real-world datasets reveal that reducing nodes can decrease computation time by up to 29%, and as a byproduct of removing noise, even improve the prediction accuracy.

关键词： Training Accuracy Graph convolutional networks Roads Noise Layout Evolutionary computation

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：