检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Murali, Nihal Puli, Aahlad Yu, Ke Ranganath, Rajesh Batmanghelich, Kayhan Intelligent Systems Program University of Pittsburgh United States Department of Computer Science New York University United States Department of Electrical and Computer Engineering Boston University United States

Deep Neural Networks (DNNs) are prone to learning spurious features that correlate with the label during training but are irrelevant to the learning problem. This hurts model generalization and poses problems when deploying them in safety-critical applications. This paper aims to better understand the effects of spurious features through the lens of the learning dynamics of the internal neurons during the training process. We make the following observations: (1) While previous works highlight the harmful effects of spurious features on the generalization ability of DNNs, we emphasize that not all spurious features are harmful. Spurious features can be "benign" or "harmful" depending on whether they are "harder" or "easier" to learn than the core features for a given model. This definition is model and dataset dependent. (2) We build upon this premise and use instance difficulty methods (like Prediction Depth (Baldock et al., 2021)) to quantify "easiness" for a given model and to identify this behavior during the training phase. (3) We empirically show that the harmful spurious features can be detected by observing the learning dynamics of the DNN’s early layers. In other words, easy features learned by the initial layers of a DNN early during the training can (potentially) hurt model generalization. We verify our claims on medical and vision datasets, both simulated and real, and justify the empirical success of our hypothesis by showing the theoretical connections between Prediction Depth and information-theoretic concepts like V-usable information (Ethayarajh et al., 2021). Lastly, our experiments show that monitoring only accuracy during training (as is common in machine learning pipelines) is insufficient to detect spurious features. We, therefore, highlight the need for monitoring early training dynamics using suitable instance difficulty metrics. © 2023, CC BY.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Multi-Robot Communication-Aware Cooperative Belief Space Planning with Inconsistent Beliefs: An Action-Consistent Approach

Multi-Robot Communication-Aware Cooperative Belief Space Pla...

引用

IEEE/RSJ International Conference on Intelligent Robots and systems (IROS)

作者： Tanmoy Kundu Moshe Rafaeli Vadim Indelman Department of Computer Science and Engineering IIIT Delhi India Technion Autonomous Systems Program (TASP) Technion - Israel Institute of Technology Haifa Israel Department of Aerospace Engineering Technion - Israel Institute of Technology Haifa Israel

ISBN: (数字)9798350377705

ISBN: (纸本)9798350377712

Multi-robot belief space planning (MR-BSP) is essential for reliable and safe autonomy. While planning, each robot maintains a belief over the state of the environment and reasons how the belief would evolve in the future for different candidate actions. Yet, existing MR-BSP works have a common assumption that the beliefs of different robots are consistent at planning time. Such an assumption is often highly unrealistic, as it requires prohibitively extensive and frequent communication capabilities. In practice, each robot may have a different belief about the state of the environment. Crucially, when the beliefs of different robots are inconsistent, state-of-the-art MR-BSP approaches could result in a lack of coordination between the robots, and in general, could yield dangerous, unsafe and suboptimal decisions. In this paper, we tackle this crucial gap. We develop a novel decentralized algorithm that is guaranteed to find a consistent joint action. For a given robot, our algorithm reasons for action preferences about 1) its local information, 2) what it perceives about the reasoning of the other robot, and 3) what it perceives about the reasoning of itself perceived by the other robot. This algorithm finds a consistent joint action whenever these steps yield the same best joint action obtained by reasoning about action preferences; otherwise, it self-triggers communication between the robots. Experimental results show efficacy of our algorithm in comparison with two baseline algorithms.

关键词： Robot kinematics Cognition Planning Reliability Robots Intelligent robots

来源：评论

学校读者我要写书评

暂无评论

Advancing Precision Aquaculture Through Big Data Analytics and Machine Learning in Canadian Fish Farming

Advancing Precision Aquaculture Through Big Data Analytics a...

引用

OCEANS 2024 - Halifax, OCEANS 2024

作者： Bravo, Francisco Amorim, Joana Amirkandeh, Melika Besharati Bodorik, Peter Cerqueira, Vitor Gomes, Nuno R.C. Korus, Jennie Oliveira, Mariana Parent, Marianne Pimentel, João Reilly, Derek Sclodnick, Tyler Grant, Jon Filgueira, Ramón Whidden, Christopher Universidad Adolfo Ibañez Facultad de Ingenieria y Ciencias Santiago Chile Dalhousie University Faculty of Computer Science Halifax Canada University of Porto Faculty of Engineering Portugal Barcelona Spain Innovasea Marine Systems Bedford Canada Charlottetown Canada Dalhousie University Department of Oceanography Halifax Canada Marine Affairs Program Dalhousie University Halifax Canada

ISBN: (纸本)9798331540081

The aquaculture industry faces significant challenges related to sustainability, productivity, and fish welfare. Key issues include managing environmental conditions, disease, pests, and data integration from various sensors and monitoring systems. The BigFish project aims to address these challenges through advanced analytics and machine learning, focusing on three case studies in Atlantic salmon farms: predicting oxygen levels, reducing sea lice infestations, and improving data interaction and visualization. Predictive models for oxygen levels and sea lice infestation, as well as natural language interfaces for data visualization, demonstrate the potential for improved decision-making and management practices in aquaculture. Early results indicate the effectiveness of these approaches, highlighting the importance of data-driven solutions in enhancing industry sustainability and productivity. © 2024 IEEE.

关键词： Aquaculture

来源：评论

学校读者我要写书评

暂无评论

AI-Powered Precision: Revolutionizing Skin Cancer Detection with Advanced Machine Learning

AI-Powered Precision: Revolutionizing Skin Cancer Detection ...

引用

Novel Intelligent and Leading Emerging Sciences Conference (NILES)

作者： Mariam M. Abd-elaziz Basma M. Elshafie Hagar Abdullah Karen Y. Gerges Ingy A. Ghanem Khaled M. Fouad Amir El-Ghamry Ibrahim F. Moawad Sara El-Metwally Faculty of computer science and engineering New Mansoura university Dakahlia Egypt Faculty of Engineering Computer and Communication program Faculty of Computer Science and Engineering Mansoura University New Mansoura University Information Systems Faculty of Computers and Artificial Intelligence Benha University Egypt Faculty of Computer and Information Sciences Faculty of Computer Science and Engineering Mansoura University New Mansoura University Egypt School of Engineering and Computer Science University of Hertfordshire GAF Egypt Faculty of Computer and Information Sciences Ain Shams University Cairo Egypt Department of Artificial Intelligence College of Computer Science and Engineering Taibah University Medina Saudi Arabia

ISBN: (数字)9798350378511

ISBN: (纸本)9798350378528

Skin cancer's increasing incidence rates necessitate advanced diagnostic tools. This research uses MobileNet architecture to develop an enhanced system for skin cancer detection. MobileNet's efficient CNN architecture, combined with Keras for model lifecycle management and TensorFlow for real-time inference, forms the foundation of this study. Using 10,015 dermatoscopic pictures from seven different classifications of skin cancer in the HAM10000 dataset, the methodology includes several key stages: image preprocessing to correct illumination and adjust resolution, data augmentation to balance the dataset, and model training using Transfer Learning. The MobileNet model was trained over 50 epochs with a comprehensive architecture incorporating multiple specialized layers. By combining large-scale data analysis and adaptive learning, this methodology demonstrates The revolutionary possibilities of AI and ML in improving skin cancer diagnostics and public health outcomes. Keywords Skin Cancer Detection, Dermatoscopic Images, Machine Learning, Deep Learning

关键词： Training Deep learning Adaptation models Image resolution Transfer learning Lighting Real-time systems Image preprocessing Public healthcare Skin cancer

来源：评论

学校读者我要写书评

暂无评论

A Constructive Heuristic for the Uniform Capacitated Vertex k-center Problem

引用

ACM Journal of Experimental Algorithmics 2023年第2期28卷 1-26页

作者： Cornejo-Acosta, José Alejandro García-Díaz, Jesús Pérez-Sansalvador, Julio César Ríos-Mercado, Roger Z. Pomares-Hernández, Saúl Eduardo Instituto Nacional de Astrofísica Óptica y Electrónica Computer Science Department Puebla72840 Mexico Tecnológico Nacional de México/ITS de Purísima Del Rincón Computer Engineering Department Guanajuato36425 Mexico Consejo Nacional de Humanidades Ciencias y Tecnologías Mexico City03940 Mexico Universidad Autónoma de Nuevo León Graduate Program in Systems Engineering San Nicolás de los GarzaNL 66455 Mexico LAAS-CNRS Université de Toulouse France INSA ToulouseF31400 France

The uniform capacitated vertex k-center problem is an -hard combinatorial optimization problem that models real situations where k centers can only attend a maximum number of customers, and the travel time or distance from the customers to their assigned center has to be minimized. This article introduces a polynomial-time constructive heuristic algorithm that exploits the relationship between this problem and the minimum capacitated dominating set problem. The proposed heuristic is based on the one-hop farthest-first heuristic that has proven effective for the uncapacitated version of the problem. We carried out different empirical evaluations of the proposed heuristic, including an analysis of the effect of a parallel implementation of the algorithm, which significantly improved the running time for relatively large instances. © 2023 Copyright held by the owner/author(s). Publication rights licensed to ACM.

关键词： Travel time

来源：评论

学校读者我要写书评

暂无评论

Distilling BlackBox to Interpretable models for Efficient Transfer Learning

arXiv

引用

arXiv 2023年

作者： Ghosh, Shantanu Yu, Ke Batmanghelich, Kayhan Department of Electrical and Computer Engineering Boston University BostonMA United States Intelligent Systems Program University of Pittsburgh PittsburghPA United States

Building generalizable AI models is one of the primary challenges in the healthcare domain. While radiologists rely on generalizable descriptive rules of abnormality, Neural Network (NN) models suffer even with a slight shift in input distribution (e.g., scanner type). Fine-tuning a model to transfer knowledge from one domain to another requires a significant amount of labeled data in the target domain. In this paper, we develop an interpretable model that can be efficiently fine-tuned to an unseen target domain with minimal computational cost. We assume the interpretable component of NN to be approximately domain-invariant. However, interpretable models typically underperform compared to their Blackbox (BB) variants. We start with a BB in the source domain and distill it into a mixture of shallow interpretable models using human-understandable concepts. As each interpretable model covers a subset of data, a mixture of interpretable models achieves comparable performance as BB. Further, we use the pseudo-labeling technique from semi-supervised learning (SSL) to learn the concept classifier in the target domain, followed by fine-tuning the interpretable models in the target domain. We evaluate our model using a real-life large-scale chest-X-ray (CXR) classification dataset. The code is available at: https://github. com/batmanlab/MICCAI-2023-Route-interpret-repeat-CXRs. © 2023, CC BY.

关键词： Large dataset

来源：评论

学校读者我要写书评

暂无评论

Enhancing Student Engagement in Distance Learning with Gamified Educational Application

Enhancing Student Engagement in Distance Learning with Gamif...

引用

International Conference on Advanced Learning Technologies (ICALT)

作者： Fernando Da Silva Pontes Luis Carlos Costa Fonseca Sofiane Labidi UEMA Graduate Program in Computer and Systems Engineering State University of Maranhão São Luís Brazil UFMA Bachelor ofScience and Technology Program -BICT Federal University of Maranhão São Luís Brazil

ISBN: (数字)9798350362053

ISBN: (纸本)9798350362060

This research explores the potential of gamified educational applications as motivational tools for distance learning students. The study draws from the research conducted by Fernando da Silva Pontes in his dissertation, which examines the impact of gamification on student motivation and engagement in the context of distance education. The paper provides a comprehensive overview of the theoretical foundations of gamification and its application in educational settings, synthesizing existing research on the impact of gamified educational applications on student motivation, engagement, and learning outcomes. The findings highlight the significant contribution of gamification in fostering a collaborative and interactive learning environment for distance education students. The study underscores the transformative potential of gamification in addressing the challenges of student motivation and engagement in distance education, paving the way for future advancements in educational practices and research. The paper provides insights into the potential of gamification in enhancing the learning experience for distance education students, contributing to the understanding of the effectiveness of gamified educational applications in motivating and engaging students in the context of distance learning.

关键词： computer aided instruction Technological innovation Education Collaboration Games Market research Stakeholders

来源：评论

学校读者我要写书评

暂无评论

Dynamic Interval-Valued PCA for Enhanced Fault Detection

Dynamic Interval-Valued PCA for Enhanced Fault Detection

引用

International Conference on Control, Decision and Information Technologies (CoDIT)

作者： Lahcene Rouani Mohamed Faouzi Harkat Abdelmalek Kouadri Abderazak Bensmail Majdi Mansouri Mohamed Nounou Signals and Systems Laboratory Institute of Electrical and Electronics Engineering University M’Hamed Bougara of Boumerdes Boumerdes Algeria Systems and Advanced Material Laboratory Badji Mokhtar - Annaba University Annaba Algeria Electrical and Computer Engineering Program Texas A&M University at Qatar Doha Qatar

ISBN: (数字)9798350373974

ISBN: (纸本)9798350373981

This study introduces three novel dynamic interval-valued principal component analysis (DIPCA) methods: dynamic centers PCA (D-CPCA), dynamic vertices PCA (D-VPCA), and dynamic complete information PCA (D-CIPCA). These methods advance traditional interval-valued PCA (IPCA) by integrating dynamic aspects of industrial processes, thus addressing both data uncertainties and temporal correlations. The DIPCA methods were validated using real-world data from the Ain El Kebira cement plant. Results indicate significant improvements in fault detection accuracy, achieving lower false alarm rates and higher reliability compared to classical IPCA methods. Furthermore, an enhanced combined index for interval-valued data was developed, providing a single, comprehensive statistical measure for streamlined process monitoring.

关键词： Process monitoring Uncertainty Accuracy Fault detection Cement industry Aerodynamics Robustness Real-time systems Indexes Principal component analysis

来源：评论

学校读者我要写书评

暂无评论

Uncertainty Quantification Kernel PCA: Enhancing Fault Detection in Interval-Valued Data

Uncertainty Quantification Kernel PCA: Enhancing Fault Detec...

引用

International Conference on Control, Decision and Information Technologies (CoDIT)

作者： Abdelhalim Louifi Abdelmalek Kouadri Mohamed Faouzi Harkat Abderazak Bensmail Majdi Mansouri Hazem Nounou Signals and Systems Laboratory Institute of Electrical and Electronics Engineering University M’Hamed Bougara of Boumerdes Boumerdes Algeria Systems and Advanced Material Laboratory Badji Mokhtar - Annaba University Annaba Algeria Electrical and Computer Engineering Program Texas A&M University at Qatar Doha Qatar

ISBN: (数字)9798350373974

ISBN: (纸本)9798350373981

The interval-valued kernel PCA (UQ-KPCA) is a variation of the kernel PCA (KPCA) designed for interval-valued data, designed to handle data uncertainty by defining specific similarity measures and kernel functions for interval data. This paper introduces Uncertainty Quantification KPCA (UQ-KPCA) as a novel method to address uncertainties in data. UQ-KPCA converts the traditional KPCA model from single-valued to interval-valued representations, allowing for accurate error and uncertainty quantification. The process modeling using KPCA is then performed on data based on the interval model, followed by the computation of fault detection statistics such as T 2 , Q, and Φ. The method’s effectiveness is evaluated in the context of the cement rotary kiln process, and compared with the KPCA demonstrating superior performance in accurately identifying faults within a stochastic setting with unknown uncertainties.

关键词： Process monitoring Kilns Uncertainty Fault detection Computational modeling Cement industry Stochastic processes Real-time systems Kernel Principal component analysis

来源：评论

学校读者我要写书评

暂无评论

Double-SP Method to Enhance Privacy and Security of Smart City Applications 10

Double-SP Method to Enhance Privacy and Security of Smart Ci...

引用

17th INDIACom;2023 10th International Conference on Computing for Sustainable Global Development, INDIACom 2023

作者： Bahbouh, Nour Mahmoud Hakami, Abdulelah Alhussain Hassan Aljohani, Yaser Mohammed Sen, Adnan Ahmed Abi Alkhodre, Ahmad Alsughayyir, Aeshah Granada University Department of Information and Communication Sciences Granada Spain Saudi Electronic University Al-Madinah Saudi Arabia YasRef Program Yanbu Saudi Arabia Computing and Information Technology Islamic University Al-Madinah Saudi Arabia Islamic University Faculty of Computer and Information Systems Al-Madinah Saudi Arabia College of Computer Science and Engineering Taibah University Al-Madinah Saudi Arabia

ISBN: (纸本)9789380544472

Smart cities promise a lot of well-being to their users in all areas of life through millions of applications and services. Smart services rely heavily on collecting data and the preferences of users. But on the dark side, the users' information is posed to threat and penetration during transmission or stored far in the clouds. Depending on encryption only is not sufficient if the attacker has strong resources or if the attacker is the service provider (SP) itself. In addition, changing data before sending is not a practical solution in many systems because of the adverse impact on the quality of a main service. This research presented a new idea to address the issue of protection. The proposed method enhanced the privacy and security of users' data without affecting the accuracy of service. The core of the suggested solution relies on a knowledge base of services that are managed by experts. Also, the solution depends on fog nodes to measure the level of security and privacy of users' queries without delay. Moreover, the fog nodes manage contact with SPs. Finally, the proposed method divided the SP into two, one for user queries and the other for user data. The simulation and analytical discussion on a practical case in smart cities demonstrated the superiority of the proposed approach over previous methods in the level of protection by maintaining the quality of services, and the resistance to attacks. © 2023 Bharati Vidyapeeth, New Delhi.

关键词： Fog

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：