检索结果-内蒙古大学图书馆

2024 Conference on Empirical Methods in Natural Language Processing, EMNLP 2024

作者： Hou, Wenjun Cheng, Yi Xu, Kaishuai Hu, Yan Li, Wenjie Liu, Jiang Department of Computing The Hong Kong Polytechnic University Hong Kong Research Institute of Trustworthy Autonomous Systems and Department of Computer Science and Engineering Southern University of Science and Technology Shenzhen China

ISBN: (纸本)9798891761681

Previous research on radiology report generation has made significant progress in terms of increasing the clinical accuracy of generated *** this paper, we emphasize another crucial quality that it should possess, i.e., inter-report consistency, which refers to the capability of generating consistent reports for semantically equivalent *** quality is even of greater significance than the overall report accuracy in terms of ensuring the system's credibility, as a system prone to providing conflicting results would severely erode users' ***, existing approaches struggle to maintain inter-report consistency, exhibiting biases towards common patterns and susceptibility to lesion *** address this issue, we propose ICON, which Improves the inter-report CONsistency of radiology report *** to enhance the system's ability to capture similarities in semantically equivalent lesions, our approach first involves extracting lesions from input images and examining their ***, we introduce a lesion-aware mixup technique to ensure that the representations of the semantically equivalent lesions align with the same attributes, achieved through a linear combination during the training *** experiments on three publicly available chest X-ray datasets verify the effectiveness of our approach, both in terms of improving the consistency and accuracy of the generated reports. © 2024 Association for Computational Linguistics.

关键词： Radiology

来源：评论

学校读者我要写书评

暂无评论

Divide-and-Conquer Strategy for Large-Scale Dynamic Bayesian Network Structure Learning 13th

Divide-and-Conquer Strategy for Large-Scale Dynamic Bayesia...

引用

13th IFIP TC 12 International Conference on Intelligent Information Processing, IIP 2024

作者： Ouyang, Hui Chen, Cheng Tang, Ke Department of Computer Science and Engineering Southern University of Science and Technology Shenzhen518055 China Research Institute of Trustworthy Autonomous Systems Southern University of Science and Technology Shenzhen518055 China

ISBN: (纸本)9783031578076

Dynamic Bayesian Networks (DBNs), renowned for their interpretability, have become increasingly vital in representing complex stochastic processes in various domains such as gene expression analysis, healthcare, and traffic prediction. Structure learning of DBNs from data is a challenging endeavor, particularly for datasets with thousands of variables. Most current algorithms for DBN structure learning are adaptations from those used in static Bayesian Networks (BNs), and are typically focused on smaller-scale problems. In order to solve large-scale problems while taking full advantage of existing algorithms, this paper introduces a novel divide-and-conquer strategy, originally developed for static BNs, and adapts it for large-scale DBN structure learning. Additionally, we leverage the prior knowledge of 2 Time-sliced BNs (2-TBNs), a special class of DBNs, to enhance the performance of this strategy. Our approach significantly improves the scalability and accuracy of 2-TBN structure learning. Designed experiments demonstrate the effectiveness of our method, showing substantial improvements over existing algorithms in both computational efficiency and structure learning accuracy. In problem instances with more than 1,000 variables, our proposed approach on average improves two accuracy metrics by 74.45% and 110.94%, respectively, while reducing runtime by an average of 93.65%. Moreover, in problem instances with more than 10,000 variables, our proposed approach successfully completed the task in a matter of hours, whereas the baseline algorithm failed to produce a reasonable result within a one-day runtime limit. © IFIP International Federation for Information Processing 2024.

关键词： Bayesian networks

来源：评论

学校读者我要写书评

暂无评论

A Comprehensive Threat Model for Autonomous Vehicle Communication Security 6

A Comprehensive Threat Model for Autonomous Vehicle Communic...

引用

6th International Conference on Advancements in Computing, ICAC 2024

作者： Meegammana, Niranjan W. Fernando, Harinda Shilpa Sayura Foundation Cyber Security Research Lab Kandy Sri Lanka Sri Lanka Institute of Information Technology Dept. of Computer Systems Engineering Malabe Sri Lanka

ISBN: (纸本)9798331517878

The rapid advancement of autonomous vehicles (AVs) promises to enhance transportation efficiency, reliability, and safety. However, these systems, particularly multi-channel vehicular communications, are vulnerable to cyber attacks that threaten authenticity, authorization, confidentiality, integrity, availability, and non-repudiation. The 2014 Chrysler Jeep hack highlights the urgent need for robust security measures in vehic-ular communications. This study utilizes the STRIDE, DREAD, and MITRE ATT&CK security frameworks to develop a comprehensive threat model that identifies and assesses critical threats and risks to AV communications. By detailing specific attack vectors and their multidimensional impacts and criticality, the study establishes a baseline of threats for communication security in vehicular networks. It produces visual threat maps, impact graphs of various attacks, and a color-coded threat grid to guide stakeholders in developing effective cyber security measures for protecting AV communications, thereby help stakeholders to improve the reliability and safety of AV navigation for fostering growth in the AV industry. © 2024 IEEE.

关键词： Cyber attacks

来源：评论

学校读者我要写书评

暂无评论

A Comprehensive Survey on Summarization Techniques

引用

SN computer Science 2023年第5期4卷 560页

作者： Uppalapati, Padma Jyothi Dabbiru, Madhavi Rao, K. Venkata Research Scholar Computer Science and Systems Engineering Department AU-TDR Hub Andhra University Andhra Pradesh Visakhapatnam 530003 India Computer Science and Engineering Department Dr.Lankapalli Bullayya College of Engineering Andhra Pradesh Visakhapatnam 530013 India Computer Science and Systems Engineering Department Andhra University Andhra Pradesh Visakhapatnam 530003 India Computer Science and Engineering Department Vishnu Institute of Technology Andhra Pradesh Bhimavaram 534202 India

Summarization is the method of extracting valuable information or a summary out of a significant amount of data using a variety of computation techniques. The summarization task has been studied for a long time. researchers are still paying attention to this task, as text can be extracted from a variety of formats, including text, audio, and multimedia, and because it presents a number of special challenges that require deep learning techniques, including temporal ordering, compression ratio, redundancy, cohesion, and coreference. In addition, the study discusses numerous cutting-edge datasets that can be used for text summarizing as well as various assessment metrics that are used to this issue. The paper’s primary goal is to summarize the various deep learning techniques currently being utilized and addresses the challenges of summarization in the education sector. © 2023, The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd.

关键词： Abstractive text summarization (ATS) Corpora Deep learning Machine learning Natural language processing (NLP) Summary evaluation metrics

来源：评论

学校读者我要写书评

暂无评论

Towards Underwater Sustainability with AIoT: Underwater Trash Management System Comprised of YOLOv8 with IoT-Applied Segmentation 2

Towards Underwater Sustainability with AIoT: Underwater Tras...

引用

2nd International Conference on Power, Electrical, Electronics and Industrial Applications, PEEIACON 2024

作者： Paneru, Biplov Poudel, Buddhi Sagar Shah, Krishna Bikram Paneru, Bishwash Chhetri, Sanjog Poudel, Yam Krishna Nepal Engineering College Pokhara University Department of Electronics and Communication Bhaktapur Nepal Nepal Engineering College Pokhara University Department of Computer Science and Engineering Bhaktapur Nepal Institute of Engineering Tribhuvan University Department of Applied Science & Chemical Engineering Lalitpur Nepal Nepal Research and Collaboration Center Department of Research and Development Kathmandu Nepal

ISBN: (纸本)9798331517984

Effective techniques for managing rubbish underwater are essential due to growing environmental concerns about contamination under the sea. In this research, we offer a novel method for efficient underwater trash detection and management that makes use of the Internet of Things (IoT) and the YOLOV8 object detection model. Using YOLOV8, the suggested system seeks to identify and categorize underwater garbage in real-time. It then uses IoT devices for segmentation and management. We outline the suggested system's design, talk about how YOLOV8 integrates with Internet of Things gadgets, and provide experimental results to validate the system's efficacy. Our findings suggest that the suggested strategy has great potential to advance efforts to manage undersea waste and manage pollution as well as segment using 6-DOF Robotic Arm. The IoT based system allows remote monitoring of trash density and location along with the raspberry-pi server-based video feed allows the users to suitably monitor aquatic conditions remotely. The fusion of IoT and deep learning on YOLO v8 model opens gateway for water sustainability and enhances further opportunities in employing robotic arm-based trash collection system. © 2024 IEEE.

关键词： Robotic arms

来源：评论

学校读者我要写书评

暂无评论

An Energy Stable Alternative WENO Scheme with a New Smoothness Indicator for Hyperbolic Conservation Laws

An Energy Stable Alternative WENO Scheme with a New Smoothne...

引用

2023 International Conference on Mechanical, Aerospace and Electronic systems, MAES 2023

作者： Lou, Shuai Ren, Yi-Peng Yang, Xue-Jun Wang, Zhao-Wei Zhong, Kang Yan, Chao Beijing Institute of Astronautical Systems Engineering Beijing100076 China Research and Development Center China Academy of Launch Vehicle Technology Beijing100076 China Shanghai Institute of Mechanical and Electrical Engineering Shanghai201100 China School of Aeronautic Science and Engineering Beihang University Beijing100191 China

An energy stable alternative weighted Essentially Non-Oscillatory (WENO) scheme with a new smoothness indicator (EAWN) is proposed in this paper. The developed EAWN scheme is stable for linear problems by adding a correction operator and can recover the designed order of accuracy at critical points with no dependency on user-defined parameter for smoothness indicators. Application of EAWN in one-dimensional and two-dimensional hyperbolic conservation laws also indicates the superiority of the new method over its non-modified version in terms of shock-capturing and vorticity-resolving. © Published under licence by IOP Publishing Ltd.

关键词： Physical properties

来源：评论

学校读者我要写书评

暂无评论

Controlled Dropout for Uncertainty Estimation

Controlled Dropout for Uncertainty Estimation

引用

2023 IEEE International Conference on systems, Man, and Cybernetics, SMC 2023

作者： Hasan, Mehedi Hossain, Ibrahim Rahman, Ashikur Nahavandi, Saeid Institute for Intelligent Systems Research and Innovation Deakin University GeelongVIC3220 Australia Bangladesh University of Engineering and Technology Dept. of Computer Science and Engineering Dhaka Bangladesh

ISBN: (纸本)9798350337020

Uncertainty quantification in a neural network is one of the most discussed topics for safety-critical applications. Though Neural Networks (NNs) have achieved state-of-the-art performance for many applications, they still provide unreliable point predictions, which lack information about uncertainty estimates. Among various methods to enable neural networks to estimate uncertainty, Monte Carlo (MC) dropout has gained much popularity in a short period due to its simplicity. In this study, we present a new version of the traditional dropout layer where we are able to fix the number of dropout configurations. As such, each layer can take and apply the new dropout layer in the MC method to quantify the uncertainty associated with NN predictions. We conduct experiments on both toy and realistic datasets and compare the results with the MC method using the traditional dropout layer. Performance analysis utilizing uncertainty evaluation metrics corroborates that our dropout layer offers better performance in most cases. © 2023 IEEE.

关键词： dropout Monte Carlo dropout neural network Uncertainty estimation

来源：评论

学校读者我要写书评

暂无评论

An Architectural Approach for Enhanced Data Interoperability Across Building systems 7

An Architectural Approach for Enhanced Data Interoperability...

引用

7th Conference on Cloud and Internet of Things, CIoT 2024

作者： Yefi, Peter Ejaz, Sikandar Menon, Ramanunni Parakkal Eicker, Ursula Guéhéneuc, Yann-Gaël Concordia University Computer Science and Software Engineering Dept. Montreal Canada Concordia University Institute for Information Systems Engineering Montreal Canada Institute for Solar Research German Aerospace Center Jülich Germany Concorida University NextGen Cities Institute Montreal Canada

ISBN: (纸本)9798350356717

Access to building data is crucial for creating portable applications and improving building operation and energy efficiency. Ensuring data transmutability is essential for facilitating research and overcoming the diverse data representation, management, and collection methods across various building management systems (BMSs). Different BMSs use multiple interfaces for data exchange. Some BMSs provide Application Programming Interfaces (APIs) for data exchanges, while others have gateways connecting to cloud services for application subscription and data access. This study introduces a five-layered architecture that describes how client applications, researchers, and other interested stakeholders can exchange data with different BMSs. Our objective is to help researchers and practitioners understand the various ways of accessing transmutable (capable of being transformed into a compatible format) data from BMSs, regardless of the building management system a building uses. Through transmutable access to building systems' data, we aim to enable transferable energy-efficient-related applications for buildings. We evaluate our work with a building model of the Varennes Library in Varennes, QC, Canada. The building model includes information about the building, floors, rooms, electric meter and historical time series data of the meter, weather station and its historical time series data. The model also has carbon dioxide concentration sensors, temperature sensors, humidity sensors, and historical time series data of the sensors. We include two test client programs to show how they access the transmutable data of the building through a model. We conclude that the five-layered architecture facilitates the exchange of transmutable data across diverse BMSs, making it a valuable tool for researchers and practitioners. © 2024 IEEE.

关键词： Middleware

来源：评论

学校读者我要写书评

暂无评论

CPDZ: A Credibility-Aware and Privacy-Preserving Data Collection Scheme with Zero-Trust in Next-Generation Crowdsensing Networks

引用

IEEE Journal on Selected Areas in Communications 2025年第6期43卷 2183-2199页

作者： Tang, Jianheng Fan, Kejia Yang, Shihao Liu, Anfeng Xiong, Neal N. Song, Houbing Herbert Leung, Victor C. M. Central South University School of Computer Science and Engineering Changsha410083 China Peking University School of Computer Science Beijing100871 China Sul Ross State University Department of Computer Mathematical and Physical Sciences AlpineTX79830 United States Department of Information Systems BaltimoreMD21250 United States Shenzhen MSU-BIT University Artificial Intelligence Research Institute Shenzhen518172 China Shenzhen University College of Computer Science and Software Engineering Shenzhen518060 China The University of British Columbia Department of Electrical and Computer Engineering VancouverBCV6T 1Z4 Canada

Next-Generation Crowdsensing Networks (NGCNs) have become increasingly critical for smart cities, where data privacy and quality are pivotal concerns. Traditional trust mechanisms in crowdsensing mainly rely on static trust models, which are insufficient for dynamic security requirements. Zero-Trust security represents a promising opportunity, yet coming with notable challenges in NGCNs, including Unknown Workers Online Recruitment (UWOR), Information Elicitation Without Verification (IEWV), Privacy Preserving Data Evaluation (PPDE), and Dynamic Trust Abrupt Shift (DTAS). To address these challenges, we propose a Credibility-aware and Privacy-preserving Data collection scheme with Zero-trust (CPDZ) for secure and quality data collection in NGCNs. First, our CPDZ scheme encompasses a quality worker recruitment strategy with combinatorial multi-armed bandit models, utilizing Thompson Sampling for the secure and efficient resolution of the UWOR. Second, an active dispatching scheme for unmanned aerial vehicles is crafted to collect data as a gold standard to assist in overcoming the IEWV challenge. Third, as for the PPDE challenge, we propose a lightweight privacy-preserving scheme for dependable truth discovery and secure trust verification. Fourth, the DTAS challenge is managed by a dual verification scheme that integrates short-term and long-term trust assessments, ensuring stability and adaptability of the zero-trust security in our CPDZ scheme. Experiments confirm the superiority of our CPDZ scheme, showing a 12.5% increase in recruitment revenue and a 57.8% reduction in relative error compared to existing approaches. © 1983-2012 IEEE.

关键词： Differential privacy

来源：评论

学校读者我要写书评

暂无评论

DHive: Query Execution Performance Analysis via Dataflow in Apache Hive 49th

DHive: Query Execution Performance Analysis via Dataflow in ...

引用

49th International Conference on Very Large Data Bases, VLDB 2023

作者： Zhang, Chaozu Shen, Qiaomu Tang, Bo Department of Computer Science and Engineering Southern University of Science and Technology China Research Institute of Trustworthy Autonomous Systems Southern University of Science and Technology China

Nowadays, Apache Hive has been widely used for large-scale data analysis applications in many organizations. Various visual analytical tools are developed to help Hive users quickly analyze the query execution process and identify the performance bottleneck of executed queries. However, existing tools mostly focus on showing the time usage of query sub-components (jobs and operators) but fail to provide enough evidence to analyze the root reasons for the slow execution progress. To tackle this problem, we develop a visual analytical system DHive to visualize and analyze the query execution progress via dataflow analysis. DHive shows the dataflow during query execution at multiple levels: query level, job level and task level, which enable users to identify the key jobs/tasks and explain their time usage by linking them to the auxiliary information such as the system configuration and hardware status. We demonstrate the effectiveness of DHive by two cases in a production cluster. © 2023, VLDB Endowment. All rights reserved.

关键词： Search engines

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：