检索结果-内蒙古大学图书馆

arXiv 2025年

作者： Long, Shaocong Zhou, Qianyu Li, Xiangtai Ying, Chenhao Tong, Yunhai Ma, Lizhuang Luo, Yuan Tao, Dacheng Department of Computer Science and Engineering Shanghai Jiao Tong University Shanghai China College of Computer Science and Technology Key Laboratory of Symbolic Computation and Knowledge Engineering Ministry of Education Jilin University Changchun China Nanyang Technological University Singapore Peking University Beijing China

Domain generalization (DG) aims to improve the generalizability of computer vision models toward distribution shifts. The mainstream DG methods predominantly focus on learning domain invariance across domains, however, such methods overlook the untapped potential inherent in domain-specific information. While the prevailing practice of discriminative linear classifier has been tailored to domain-invariant features, it struggles when confronted with diverse domain-specific information, e.g., intra-class shifts, that exhibits multi-modality. To address these issues, we explore the theoretical implications of relying on domain-invariant features, revealing the crucial role of domain-specific information in mitigating the target risk for DG. Drawing from these insights, we propose Generative Classifier-driven Domain Generalization (GCDG), introducing a generative paradigm for the DG classifier based on Gaussian Mixture Models (GMMs) for each class across domains. GCDG consists of three key modules: Heterogeneity Learning Classifier (HLC), Spurious Correlation Blocking (SCB), and Diverse Component Balancing (DCB). Concretely, HLC attempts to model the feature distributions and thereby capture valuable domain-specific information via GMMs. SCB identifies the neural units containing spurious correlations and perturbs them, mitigating the risk of HLC learning irrelevant spurious patterns. Meanwhile, DCB ensures a balanced contribution of components within HLC, preventing the underestimation or neglect of critical components. In this way, GCDG excels in capturing the nuances of domain-specific information characterized by diverse distributions. GCDG demonstrates the potential to reduce the target risk and encourage flat minima, improving the model’s generalizability. Extensive experiments show GCDG’s comparable performance on five DG benchmarks and one face anti-spoofing dataset, seamlessly integrating into existing DG methods with consistent improvements. Code will be avail

关键词： Discriminant analysis

来源：评论

学校读者我要写书评

暂无评论

Analyzing P300 Signals to Assess Undergraduate Attention Span Using Impulse Neiry BCI

Analyzing P300 Signals to Assess Undergraduate Attention Spa...

引用

Brain computer Interface & Healthcare Technologies (iCon-BCIHT), International Conference on

作者： Kavitha Sooda Indiramma M Sanjeev Kubakaddi Department of Computer Science and Engineering B.M.S. College of Engineering Bangalore India Department of AI and DS B.M.S. College of Engineering Bangalore India itie Knowledge Solutions Bangalore India

ISBN: (数字)9798331540067

ISBN: (纸本)9798331540074

The study on brain signals to understand working memory in undergraduate students has yielded promising results. However, the challenge remains in accurately assessing different brain activities. This work seeks to measure concentration in cognitive performance, offering a detailed dataset that tracks how individuals respond to cognitive tasks. A virtual reality-based solution has proven effective in capturing the P300 signal, providing a more focused and accurate measurement platform. The study aims to evaluate the efficacy of P300 signals as indicators of attention and cognitive engagement in educational settings. A brain-computer interface system, Impulse Neiry, captured and analyzed P300 signals from undergraduate students while performing various cognitive tasks. User attention span was evaluated by measuring parameters such as session duration and interaction frequency with the software, offering insights into the system's usability. The results indicated that participants experienced about 35% deeper relaxation, and their attention spans increased by 18.97% compared to baseline engagement levels.

关键词： Headphones Medical services Games Particle measurements Software Electroencephalography Frequency measurement Cognitive science Software measurement Usability

来源：评论

学校读者我要写书评

暂无评论

Advanced Deep Reinforcement Learning Strategies for Enhanced Autonomous Vehicle Navigation Systems

Advanced Deep Reinforcement Learning Strategies for Enhanced...

引用

computer, Communication and Control (IC4), International Conference on

作者： M Dhinakaran R.Thalapathi Rajasekaran V. Balaji V. Aarthi S. Ambika Department of Electronics and Communication Engineering Galgotias College of Engineering & Technology Knowledge Park II Greater Noida Uttar Pradesh India Department of Computer Science and Engineering Saveetha School of Engineering Saveetha Institute of Medical and Technical Sciences Saveetha University Chennai India Department of Computer Science and Engineering Easwari Engineering College Chennai India Department of Computer Science and Engineering GRT Institute of Engineering and Technology Tiruttani Tamil Nadu India Department of Information Technology E.G.S. Pillay Engineering College Nagapattinam Tamil Nadu India

The objective of this research is to investigate more sophisticated techniques in Deep Reinforcement Learning (DRL) for the purpose of improving navigation systems for autonomous vehicles. The results of our research indicate that the success rates of DRL algorithms are encouraging: DQN (75.2%), PPO (82.6%), and TRPO (78.1%). The success of navigation is highly influenced by optimisation factors such as parameter tuning (87.4%), exploration-exploitation balance (85.1%), and reward shaping (88.9%). The generalisation score for transfer learning is rather good (92.3%), while the safety modules are designed to make the system safer. These findings highlight the significance of algorithm selection, parameter optimisation, and advanced techniques in the process of optimising navigation for autonomous vehicles, hence establishing essential benchmarks for future developments.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Secure Pool Mining Through SVM-Based Miner Classification and Computation Validation in Blockchain Networks

Secure Pool Mining Through SVM-Based Miner Classification an...

引用

Information Technology (OCIT), OITS International Conference on

作者： Srinuvas Kumar Battina Ravi Kumar Suggala Srinivasa Rao Dangeti Meriga Kiran Kumar Pavan Kumar Vadrevu Naga Sravanthi Puppala Information Technology SR Gudlavalleru Engineering College Gudlavalleru AP Seshadri Rao Knowledge Village India Department of Information Technology Shri Vishnu Engineering College for Women Bhimavaram India Department of Computer Science and Engineering Lakireddy Bali Reddy College of Engineering Mylavaram India Department of Computer Science and Engineering School of Engineering and Applied Sciences SRM University AP Amaravati India

ISBN: (数字)9798331510404

ISBN: (纸本)9798331510411

Since the adoption of the Blockchain network by modern technologies, which benefit from its immutability and secure access provided by its consensus process, blockchain mining has gained popularity. Blockchain networks' miners rely on the consensus procedure to contribute data to the network. Miners use their resources to publish new blocks and receive rewards, and they participate in the mining process. Since solo mining might be difficult, most miners decided to join a mining pool. Several issues surrounding the miner's choice with pool mining remain unanswered because of the range of many mining pools and the potential adoption of various incentive systems. There are a few issues to consider, though, about maintaining a miner's trust throughout the consensus and mining processes. As part of the consensus, miners that are part of a mining pool contribute their processing power toward the task of adding a block. The pool will be rewarded if they are successful in their endeavors. To guarantee the execution of the consensus protocol in pool mining, a trust evaluation model is presented in our study. As a result, trust assessment among miners is formulated in this research it is evaluated as a classification problem, and an innovative approach using machine learning is offered to classify trust. First, the vector for the trust features is built using the parameters of the Blockchain. A trust classifier can then be created by training it with samples of data that miners have acquired and that contain transaction activities. Also, the provided Trust model verifies that the total computing power of the pool for which miners are requested and the miner's computation power should be 50 percent greater than or equal to the total computation power this is for to mitigate attacks brought on by untrusted miners through forks in the Blockchain. The Hyperledger tool was used as the framework to assure robustness in analyzing transaction delay, and transaction throughput, and to

关键词： Training Distributed ledger Computational modeling Machine learning Throughput Vectors Robustness Blockchains Data mining Security

来源：评论

学校读者我要写书评

暂无评论

Drawing Order Diagrams Through Two-Dimension Extension

引用

Journal of Graph Algorithms and Applications 2023年第9期27卷 783-802页

作者： Dürrschnabel, Dominik Hanika, Tom Stumme, Gerd Knowledge & Data Engineering Group Research Center for Information System Design & Department of Electrical Engineering and Computer Science University of Kassel Germany Institute of Computer Science University of Hildesheim Germany Berlin School of Library and Information Science Humboldt-Universität zu Berlin Germany

Ordinal real-world data such as concept hierarchies, ontologies, genealogies, or task dependencies in scheduling often has the property to not only contain pairwise comparable, but also incomparable elements. Order diagrams provide an important tool for understanding and thus discovering knowledge in such data. Easily readable drawings of such order diagrams are hard to come by, even for small ordered sets. Many attempts were made to transfer classical graph drawing approaches to order diagrams. Although these methods produce satisfying results for some ordered sets, they unfortunately perform poorly in general. In this work, we present the novel algorithm DimDraw to decompose an ordered set (e.g., a concept hierarchy) in linear orders and to produce a corresponding order diagram. This algorithm is based on a relation between the dimension of an ordered set and the bipartiteness of its transitive incompatibility graph. To evaluate the quality of the algorithm, a user study was conducted where generated drawings were compared with ones from state-of-the-art drawing algorithms. © 2023 Brown University. All rights reserved.

关键词： Set theory

来源：评论

学校读者我要写书评

暂无评论

ANFIS-based Multilayered Algorithm for Botnet Detection

ANFIS-based Multilayered Algorithm for Botnet Detection

引用

Recent Advances in Electrical, Electronics, Ubiquitous Communication, and Computational Intelligence (RAEEUCCI), International Conference on

作者： T Thamaraimanalan C Venkatesan M Ramkumar A Sivaramakrishnan M Marimuthu Department of Electronics and Communication Engineering Sri Eshwar College of Engineering Coimbatore India Department of Electronics and Communication Engineering HKBK College of Engineering Bangalore India Department of Computer Science and Business System Knowledge Institute of Technology Salem India Department of Computer Science and Engineering Koneru Lakshmaiah Education Foundation Andhra pradesh India Department of Computer Science and Engineering Sona College of Technology Salem India

In this study, a multilayered botnet detection system for network-based botnet detection based on ANFIS is suggested. Two layers make up the suggested system, one of which is utilized for feature extraction and the other for botnet detection. Using mean and variance, the feature extraction layer pulls features from network traffic data. The botnet detection layer employs ANFIS to identify botnet activity based on the collected features. Using the CICIDS2017 dataset, the suggested system's performance is assessed. The experimental findings demonstrate that the suggested algorithm works better than other current botnet detection techniques regarding accuracy, precision, and recall. The proposed system obtained 99.83% accuracy, 99.75% precision, and 99.91 % recall. The proposed multilayered ANFIS-based botnet detection system offers an excellent way to identify botnet activity in a network. It can be utilized in various network security applications, such as threat intelligence, malware detection, and intrusion detection.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Machine Learning-Driven Blockchain for Enhanced Drug Discovery and Development in Pharmaceutical Research

Machine Learning-Driven Blockchain for Enhanced Drug Discove...

引用

2024 OPJU International Technology Conference on Smart Computing for Innovation and Advancement in Industry 4.0, OTCON 2024

作者： Asha, V. Anjimoon, Shaik Verma, Meenu Rani Singla, Atul Khan, Irfan Alkhafaji, Mohammed Ayad New Horizon College of Engineering Department of Computer Applications Bangalore India Institute of Aeronautical Engineering Hyderabad Dundigal India GyanGanga Institute of Technology and Sciences DoCSBS M.P Jabalpur India Lovely Professional University Phagwara India Lloyd Institute of Engineering & Technology Knowledge Park II Uttar Pradesh Greater Noida201306 India National University of Science and Technology College of Technical Engineering Dhi Qar64001 Iraq

ISBN: (纸本)9798350373783

Machine learning (ML) and blockchain technologies have made huge steps forward in the pharmacy business by being used in the process of creating and discovering new medicines. Both breakthroughs were made possible by using both tools together. With the help of a blockchain design powered by machine learning, the study's main goal is to come up with a new way to make drug research more effective and efficient. This method makes it possible to make molecules with a wide range of structure features that have the right pharmacological qualities. You can also make these molecules, improve the plans of clinical trials, and guess how drugs will interact with their targets. Technologies like generative adversarial networks, reinforcement learning methods, and deep learning algorithms have made it possible for these things to be used. If blockchain technology is used throughout the whole process of making medicines, it improves global openness, traceability, and data security. This is because it makes sure that managing data is safe. The tests show that the suggested method works better than the others when it comes to the process of finding and developing new drugs. This is better because the creation process is more precise, accurate, and efficient. © 2024 IEEE.

关键词： Drug discovery

来源：评论

学校读者我要写书评

暂无评论

Advancing Wire Arc Additive Manufacturing: A New Era of Precision and Efficiency in Large Metal Component Production

Advancing Wire Arc Additive Manufacturing: A New Era of Prec...

引用

2024 OPJU International Technology Conference on Smart Computing for Innovation and Advancement in Industry 4.0, OTCON 2024

作者： Khan, Irfan Anjimoon, Shaik Asha, V. Singla, Atul Akhil, P. Alsultany, Nahed Mahmood Ahmed Lloyd Institute of Engineering & Technology Knowledge Park II Uttar Pradesh Greater Noida201306 India Institute of Aeronautical Engineering Hyderabad Dundigal India New Horizon College of Engineering Department of Computer Applications Bangalore India Lovely Professional University Phagwara India St. Martin's Engineering College Dept. of Cse Telangana Secunderabad India National University of Science and Technology Collage of Dentist Dhi Qar64001 Iraq

ISBN: (纸本)9798350373783

A comprehensive Wire Arc Additive Manufacturing (WAAM) approach that employs five algorithms to improve massive metal components is described in this research. The adaptive layer thickness control algorithm adjusts layer thickness based on real-time complexity. This ensures additive manufacturing accuracy throughout. Also, the Material Versatility Optimization Algorithm simplifies working with multiple materials, making casting more efficient. The Intelligent Path Planning Algorithm allocates the robotic arm's movement to traverse the shortest distance and uniformly distribute heat during deposition. The MultiMaterial-Materialon Strategy Algorithm improves material amounts, making WAAM more usable for additional materials. The Real-Time Excellent Assurance Algorithm constantly examines and adjusts parameters to ensure excellent production. A complete analysis of research reveals how algorithms are related and how vital they are for accuracy, speed, and flexibility. Testing shows that the recommended technique controls layer thickness, material variety, deposition velocity, accuracy, energy efficiency, and real-time tracking better than typical WAAM methods. Figures 6, 7, and 8 indicate that the proposed method delivers superior results across several criteria. The recommended WAAM approach can create more massive metal components by fundamentally changing how things are done. This research provides a comprehensive additive manufacturing approach to solve numerous large-scale metal production issues. © 2024 IEEE.

关键词： Quality control

来源：评论

学校读者我要写书评

暂无评论

Optimizing Heart Attack Prediction Through OHE2LM: A Hybrid Modelling Strategy

引用

Journal of Electrical Systems 2024年第1期20卷 66-75页

作者： Mall, Pawan Kumar Srivastava, Swapnita Patel, Mitul M. Kumar, Aniruddh Narayan, Vipul Kumar, Sanjay Singh, P.K. Singh, D.S. GL Bajaj Institute of Technology and Management India Department of Electronics and Communication Engineering Parul Institute of Engineering and Technology Parul University Vadodara India Department of Computer Science and Engineering Galgotias College of Engineering and Technology Knowledge Park-2 Greater Noida India Galgotias University Uttar Pradesh Gautam Buddha Nagar India Rajkiya Engineering College Azamgarh India Computer Science and Engineering Department Madan Mohan Malaviya University Of Technology Uttar Pradesh Gorakhpur273010 India Computer Science and Engineering Department Madan Mohan Malaviya University of Technology Uttar Pradesh Gorakhpur273010 India

Predicting heart attacks stands as a significant concern contributing to global morbidity. Within clinical data analysis, cardiovascular disease emerges as a pivotal focus for forecasting, wherein Data science and machine learning (ML) offer invaluable tools. These methodologies aid in predicting heart attacks by considering various risk factors Just like high blood pressure, increased cholesterol levels, irregular pulse rates, and diabetes, this research aims to enhance the accuracy of predicting heart disease through machine learning *** study introduces a MLdriven approach, termed ML-ELM, dedicated to forecasting heart attacks by analysing diverse risk factors. The proposed ML-ELM model is compared with alternative Utilizing machine learning techniques like Support Vector Machines, Logistic Regression, Naïve Bayes, and XGBoost is a key aspect of this exploration into different approaches for predictive modeling., is part of the research strategy. The dataset utilized for heart disease symptoms is sourced from the UCI ML Repository. The outcomes reveal that our proposed ML-ELM model has demonstrated superior predictive performance among the ML techniques tested. ML models show notable efficiency in identifying heart attack symptoms, particularly with boosting algorithms. Accuracy assessments were employed to gauge the predictive ability, Our suggested model demonstrated an outstanding accuracy rate of 96.77%. © JES 2024 on-line: ***

关键词： Support vector machines

来源：评论

学校读者我要写书评

暂无评论

PSDNet: A Breakthrough Parking Space Detection Network Powered by YOLOv8

PSDNet: A Breakthrough Parking Space Detection Network Power...

引用

Advances in Computing, Communication and Applied Informatics (ACCAI), International Conference on

作者： Prabu Selvam P. Saravanan M. Marimuthu M. R. Nithya V. Sathiyapriya V. Nareen Harsha Department of Computer Science and Engineering SRM Institute of Science & Technology Tiruchirappalli India School of Computer Science and Engineering VIT University Chennai India Dept. of CSBS K. Ramakrishnan College of Engineering Trichy India Department of CSE Knowledge Institute of Technology Salem India Dept. of CSE (AIE) Amrita VishwaVidyapeetham Chennai India

ISBN: (数字)9798350389449

ISBN: (纸本)9798350389456

Parking congestion presents a significant challenge in urban areas worldwide, spanning both developed and developing nations. The escalating volume of vehicles compounds the scarcity of parking spots, particularly in downtown areas and adjacent streets. Local authorities are tasked with enforcing regulations to mitigate this issue, yet the current state of affairs often falls short of meeting the needs of numerous residents. Consequently, this article advocates using the YOLOv8 algorithm to identify vacant parking spaces. The proposed YOLOv8 comprises three pivotal blocks: the Backbone, Neck, and Head. The Backbone segment employs an enhanced version of the CSPDarknet53 network, integrating cross-stage partial connections to strategically strengthen information exchange across its diverse layers. The Neck section merges features with varying scales and incorporates contextual information from the feature maps. The Head section predicts class, objectness scores, and bounding boxes of identified objects. The performance of the YOLOv8 algorithm in identifying vacant parking spaces is evaluated using the parking space availability dataset. It achieves an F1-score of $98.7 \%$, which is $8.7 \%$ higher than the existing object detector.

关键词： Local government Urban areas Detectors Prediction algorithms Regulation Neck Object recognition

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：