检索结果-内蒙古大学图书馆

Pattern Recognition and Artificial Intelligence (PRAI), International Conference on

作者： Ammara Sattar Muhammad Assam Tahani Jaser Alahmadi Uzair Aslam Bhatti Hao Tang Muhammad Aamir Harbin Engineering University Harbin China Department of Software Engineering University of Science and Technology Bannu KPK Pakistan College of Computer and Information Sciences Princess Nourah bint Abdulrahman University Riyadh Saudi Arabia School of information and communication engineering Hainan University Haikou China College of Computer Science Huanggang Normal University Huanggang China

ISBN: (数字)9798350350890

ISBN: (纸本)9798350350906

In the domain of image description generation, the focal point mechanism is crucial for highlighting the most relevant features within an image, thereby influencing the quality of generated descriptions. To address this challenge, we introduce an innovative optimization module called the Enhanced Focal (EF) mechanism, which builds upon the traditional multi-head focal point mechanism. The EF module is designed to assess the correlation levels between objects within an image, thereby guiding the process of subtitle generation more effectively. The EF mechanism integrates the output of the multi-head focal point mechanism with the current context (i.e., the query) to generate two components: an “information vector” and an “attention gate.” The “attention gate” is then applied to the “information vector” to create a refined focal point mechanism. This refined mechanism undergoes element-wise multiplication with the “information vector,” producing attention information that reveals a strong correlation between the target and the background in the image. This process leverages the Long Short-Term Memory (LSTM) mechanism to enhance the model's capability to capture intricate relationships. Upon training and evaluating our enhanced model on the MS COCO dataset, we observed significant improvements in key evaluation metrics. The BLEU-1 and METEOR scores reached 73.01% and 28.81%, respectively, showcasing the effectiveness of the proposed EF mechanism in improving image description generation.

关键词： Training Correlation Accuracy Logic gates Vectors Stability analysis Pattern recognition Long short term memory Optimization Remote sensing

来源：评论

学校读者我要写书评

暂无评论

Learning Shape-Preserving Autoencoder for the Reconstruction of Functional Data from Noisy Observations 23rd

Learning Shape-Preserving Autoencoder for the Reconstructi...

引用

23rd International Conference on Computational Science, ICCS 2023

作者： Krzyżak, Adam Rafajlowicz, Wojciech Rafajlowicz, Ewaryst Department of Computer Science and Software Engineering Concordia University 1455 De Maisonneuve Blvd. West MontrealQCH3G 1M8 Canada Szczecin Poland Faculty of Information and Communication Technology Wroclaw University of Science and Technology Wroclaw Poland

ISBN: (纸本)9783031360206

We propose a new autoencoder preserving input functions’ general shape (monotonicity, convexity) after their reconstruction without imposing a priori constraints. These properties are inherent to the coefficients of the Bernstein-Durrmeyer polynomials that serve here as theoretical descriptors. Their estimates, computed from noisy observations by the coder, play the role of latent variables. The approach is purely nonparametric, i.e., no prior finite-dimensional model is assumed. The answer to the question of how many latent variables should be used for an acceptable reconstruction accuracy of a family of functional data is inferred from learning based on the proposed approximation of the Akaike Information Criterion. A distinguishing feature of this autoencoder is that the coder and encoder are designed as precomputed and stored matrices with the Bernstein polynomial entries. Thus, after selecting the number of latent variables, the autoencoder usage has a low computational complexity since it is linear with respect to observations. The proposed computational algorithms are tested on real data arising in mechanical engineering when control of damping vibrations is necessary. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Polynomials

来源：评论

学校读者我要写书评

暂无评论

Virtual Reality-Based Preoperative Planning for Optimized Trocar Placement in Thoracic Surgery: A Preliminary Study

arXiv

引用

arXiv 2024年

作者： Harirpoush, Arash Rakovich, George Kersten-Oertel, Marta Xiao, Yiming Department of Computer Science and Software Engineering Concordia University Quebec Canada Maisonneuve Rosemont Hospital University of Montreal Quebec Canada School of Health Concordia University Quebec Canada

Video-assisted thoracic surgery (VATS) is a minimally invasive approach for treating early-stage non-small-cell lung cancer. Optimal trocar placement during VATS ensures comprehensive access to the thoracic cavity, provides a panoramic endoscopic view, and prevents instrument crowding. While established principles such as the Baseball Diamond Principle (BDP) and Triangle Target Principle (TTP) exist, surgeons mainly rely on experience and patient-specific anatomy for trocar placement, potentially leading to sub-optimal surgical plans that increase operative time and fatigue. To address this, we present the first virtual reality (VR)-based pre-operative planning tool with tailored data visualization and interaction designs for efficient and optimal VATS trocar placement, following the established surgical principles and consultation with an experienced surgeon. In our preliminary study, we demonstrate the system’s application in right upper lung lobectomy, a common thoracic procedure typically using three trocars. A preliminary user study of our system indicates it is efficient, robust, and user-friendly for planning optimal trocar placement, with a great promise for clinical application while offering potentially valuable insights for the development of other surgical VR systems. © 2024, CC BY-NC-ND.

关键词： Lung cancer

来源：评论

学校读者我要写书评

暂无评论

CASCADING BANDITS ROBUST TO ADVERSARIAL CORRUPTIONS

arXiv

引用

arXiv 2025年

作者： Xie, Jize Chen, Cheng Wang, Zhiyong Li, Shuai Department of Industrial Engineering Hong Kong University of Science and Technology Hong Kong Software Engineering Institute East China Normal University China Department of Computer Science The Chinese University of Hong Kong Hong Kong John Hopcroft Center Shanghai Jiao Tong University China

Online learning to rank sequentially recommends a small list of items to users from a large candidate set and receives the users’ click feedback. In many real-world scenarios, users browse the recommended list in order and click the first attractive item without checking the rest. Such behaviors are usually formulated as the cascade model. Many recent works study algorithms for cascading bandits, an online learning to rank framework in the cascade model. However, the performance of existing methods may drop significantly if part of the user feedback is adversarially corrupted (e.g., click fraud). In this work, we study how to resist adversarial corruptions in cascading bandits. We first formulate the "Cascading Bandits with Adversarial Corruptions" (CBAC) problem, which assumes that there is an adaptive adversary that may manipulate the user feedback. Then we propose two robust algorithms for this problem, which assume the corruption level is known and agnostic, respectively. We show that both algorithms can achieve logarithmic regret when the algorithm is not under attack, and the regret increases linearly with the corruption level. The experimental results also verify the robustness of our methods. Copyright © 2025, The Authors. All rights reserved.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

A Deep Learning Approach for Face Detection using Max Pooling 5

A Deep Learning Approach for Face Detection using Max Poolin...

引用

5th International Conference on Trends in Electronics and Informatics, ICOEI 2021

作者： Mehedi Shamrat, F.M. Javed Jubair, Md. Al Billah, Md. Masum Chakraborty, Sovon Alauddin, Md. Ranjan, Rumesh Daffodil International University Department of Software Engineering Dhaka Bangladesh European University of Bangladesh Department of Computer Science and Engineering Dhaka Bangladesh Department of Plant Breeding and Genetics Punjab Agriculture University Punjab India

ISBN: (纸本)9781665415712

Deep learning is a trendy term these days, and it refers to a modern age in machine learning in which algorithms are taught to identify patterns in vast amounts of data. It mostly refers to studying various layers of representation, which assists in the understanding of data that includes text, sound, and pictures. To interact with the objects in a video series, many researchers use a form of deep learning called a CNN. Face detection involves several face-related technologies, such as face authentication, facial recognition, and face clustering. For identification and understanding, effective preparation must be carried out. The standard technique did not produce a positive outcome in terms of face recognition precision. The objectives of this research are by using a deep learning model to enhance the accuracy of face detection. For recognizing faces from datasets, the proposed model utilizes a deep learning technique named convolutional neural networks. The proposed work is applied using Max Pooling, a well-known deep learning process. Our model is trained and validated using the LFW dataset, which includes 13000 photos collected from Kaggle. The training accuracy of the model was 95.72% percent, and the validation accuracy was 96.27%. © 2021 IEEE.

关键词： Face recognition

来源：评论

学校读者我要写书评

暂无评论

A column generation algorithm with dynamic constraint aggregation for minimum sum-of-squares clustering

arXiv

引用

arXiv 2024年

作者： Sudoso, Antonio M. Aloise, Daniel Department of Computer Control and Management Engineering "Antonio Ruberti" Sapienza University of Rome Via Ariosto 25 Rome00185 Italy Department of Computer and Software Engineering Polytechnique Montréal 2500 Chem. de Polytechnique MontréalH3T 1J4 Canada

The minimum sum-of-squares clustering problem (MSSC), also known as k-means clustering, refers to the problem of partitioning n data points into k clusters, with the objective of minimizing the total sum of squared Euclidean distances between each point and the center of its assigned cluster. We propose an efficient algorithm for solving large-scale MSSC instances, which combines column generation (CG) with dynamic constraint aggregation (DCA) to effectively reduce the number of constraints considered in the CG master problem. DCA was originally conceived to reduce degeneracy in set partitioning problems by utilizing an aggregated restricted master problem obtained from a partition of the set partitioning constraints into disjoint clusters. In this work, we explore the use of DCA within a CG algorithm for MSSC exact solution. Our method is fine-tuned by a series of ablation studies on DCA design choices, and is demonstrated to significantly outperform existing state-of-the-art exact approaches available in the literature. Copyright © 2024, The Authors. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Runtime Backdoor Detection for Federated Learning via Representational Dissimilarity Analysis

引用

IEEE Transactions on Dependable and Secure Computing 2025年

作者： Zhang, Xiyue Xue, Xiaoyong Du, Xiaoning Xie, Xiaofei Liu, Yang Sun, Meng University of Oxford Department of Computer Science United Kingdom Peking University Department of Information and Computing Science School of Mathematical Science China Monash University Department of Software Systems and Cybersecurity Australia Singapore Management University School of Computing and Information Systems Singapore Nanyang Technological University School of Computer Science and Engineering Singapore

Federated learning (FL), as a powerful learning paradigm, trains a shared model by aggregating model updates from distributed clients. However, the decoupling of model learning from local data makes FL highly vulnerable to backdoor attacks, where a single compromised client can poison the shared model. While recent progress has been made in backdoor detection, existing methods face challenges with detection accuracy and runtime effectiveness, particularly when dealing with complex model architectures. In this work, we propose a novel approach to detecting malicious clients in an accurate, stable, and efficient manner. Our method utilizes a sampling-based network representation method to quantify dissimilarities between clients, identifying model deviations caused by backdoor injections. We also propose an iterative algorithm to progressively detect and exclude malicious clients as outliers based on these dissimilarity measurements. Evaluations across a range of benchmark tasks demonstrate that our approach outperforms state-of-the-art methods in detection accuracy and defense effectiveness. When deployed for runtime protection, our approach effectively eliminates backdoor injections with marginal overheads. © 2025 IEEE.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Hybrid Makes Better: Privacy-Preserving Medical Image Classification Based on Federated Learning

Hybrid Makes Better: Privacy-Preserving Medical Image Classi...

引用

IEEE Conference on Global Communications (GLOBECOM)

作者： Zhi Wang Haotian Chi Yonggang Li Yuxuan Zhang Shunrong Jiang Xiaojiang Du Mohsen Guizani School of Computer Science China University of Mining and Technology Xuzhou China School of Automation and Software Engineering Shanxi University Taiyuan China Department of Electrical and Computer Engineering Stevens Institute of Technology Hoboken NJ USA Department of Machine Learning Mohamed Bin Zayed University of Artificial Intelligence Abu Dhabi UAE

ISBN: (数字)9798350351255

ISBN: (纸本)9798350351262

The power of machine learning makes it available for medical image classification. However, machine learning requires large medical datasets to improve accuracy, which will involve patients’ private information and lead to their privacy leakage. Federated learning is a trending technique to both protect patients’ privacy and improve the accuracy of medical image classification. Unfortunately, current research shows that federated learning faces the risk of privacy leakage. In this paper, we propose a privacy-preserving scheme FHDM. Specifically, we construct a local hybrid differential privacy algorithm (LHDP) against patients’ privacy leakage. This hybrid algorithm utilizes both Gaussian and Laplace differential privacy without enlarging the privacy budget. We prove that the algorithm applies to the model parameter. Moreover, we design loss optimization and global optimization strategies on the algorithm to achieve higher accuracy in medical image classification. Finally, we validate FHDM in terms of privacy-preserving and model accuracy on real datasets. Experiments show that FHDM effectively protects privacy and improves the average accuracy by 10.0% compared to adopting the LDP-based scheme with the same medical image dataset and privacy budget.

关键词： Privacy Differential privacy Accuracy Federated learning Inference algorithms Classification algorithms Security Optimization Biomedical imaging Image classification

来源：评论

学校读者我要写书评

暂无评论

AI-Enhanced Multi-Cloud Security Management: Ensuring Robust Cybersecurity in Hybrid Cloud Environments

AI-Enhanced Multi-Cloud Security Management: Ensuring Robust...

引用

Innovative Computing, Intelligent Communication and Smart Electrical Systems (ICSES), International Conference on

作者： Srimathi. J K. Kanagasabapathi Kirti Mahajan Shahanawaj Ahamad E. Soumya Shivangi Barthwal Department of Information Technology KPR College Of Arts Science And Research Arasur Covai Computer Science and Technology Karpagam College of Engineering Coimbatore Department of Information Technology Institute of Management & Entrepreneurship Development Bharati Vidyapeeth Deemed University Pune Department of Software Engineering College of Computer Science and Engineering University of Hail Hail City Saudi Arabia Department of Computer Science and Engineering St. Martins Engineering College Secunderabad Telangana Department of Electronics & Communication Engineering Graphic Era Deemed to be University Dehradun Uttarakhand India

The complex landscape of multi-cloud settings is the result of the fast growth of cloud computing and the ever-changing needs of contemporary organizations. Strong cyber defenses are of fundamental importance in this setting. In this study, we investigate the use of AI in hybrid cloud settings for the purpose of multi-cloud security management. To help businesses improve their productivity and resilience, we provide a mathematical model for optimal resource allocation. Our methodology streamlines dynamic threat assessments, making it easier for security teams to efficiently priorities vulnerabilities. The advent of a new age of real-time threat response is heralded by the incorporation of AI-driven security tactics. The technique we use has real-world implications that may help businesses stay ahead of constantly changing threats. In the future, scientists will focus on autonomous security systems, interoperability, ethics, interoperability, and cutting-edge AI models that have been validated in the real world. This study provides a detailed road map for businesses to follow as they navigate the complex cybersecurity landscape of multi-cloud settings, therefore promoting resilience and agility in this era of digital transformation.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Dynamic Selection Scheme of LTE-M and 5G for Efficient Maritime Communication Based on Traffic Volume 24

Dynamic Selection Scheme of LTE-M and 5G for Efficient Marit...

引用

24th Asia-Pacific Network Operations and Management Symposium, APNOMS 2023

作者： Yoon, Mahnsuk Cho, Kyucheol Kwon, Jaeuk Seo, Jihun Lee, Hyunjoo Yun, Jangkyu Gumi Korea Republic of School of Computer Secience and Engineering Software Education Center Kyungpook National University Daegu Korea Republic of Korea Transport Institute Center for Connected & Automated Driving Research Department of The Fourth Industrial Revolution & Transport Sejong-si Korea Republic of Gumi Korea Republic of

ISBN: (纸本)9788995004395

This paper investigates the technology for selecting communication scheme based on the amount of traffic in LTE-M (Marine) and 5G communication environments. Recently, it has been possible to use LTE-M and 5G communications for preventing collisions between ships and obstacles in coastal areas, moving from a focus on terrestrial users. This paper proposes a scheme for efficiently handling both small-scale and large-volume traffic generated from terminals on ships. LTE-M can accommodate more terminals by utilizing the wide coverage of coastal areas, but it can be vulnerable to high-volume video processing for real-time obstacle detection. 5G provides a high data transmission rate and low latency, making it suitable for large data exchanges between ships or real-time services, but its short coverage makes it unsuitable for accommodating a wide area. Therefore, this paper proposes a scheme for dynamically selecting LTE-M or 5G communication schemes according to the amount of data generated by terminals on ships, to transmit data. In this paper, we propose an algorithm to decide which communication scheme, LTE-M or 5G, should be chosen based on the amount and type of traffic. The algorithm dynamically selects the optimal communication scheme considering real-time traffic information and network status. By implementing a simulation environment for coastal areas using the technique proposed in this paper, we compare the scheme where terminals installed on ships selectively use LTE-M and 5G communication schemes with the scheme where two communication schemes are chosen randomly for transmission. The simulation results confirm that the proposed technique can process more data compared to the existing schemes. Copyright 2023 KICS.

关键词： Ships

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：