检索结果-内蒙古大学图书馆

Exploiting the Potential of Self-Supervised Monocular Depth Estimation via Patch-Based Self-Distillation

IEEE Internet of Things Journal 2025年第11期12卷 18042-18055页

作者： Yu, Ting Pan, Shan Chen, Wei Tian, Zijian Wang, Zehua Yu, F. Richard Leung, Victor C. M. Postdoctoral Research Station China RongTong Academy of Sciences Group Corporation Limited Beijing100022 China China RongTong Artificial Intelligence Research Center Beijing100022 China China University of Mining and Technology School of Artificial Intelligence Beijing100083 China China University of Mining and Technology School of Computer Science and Technology Xuzhou221116 China School of Artificial Intelligence Beijing100083 China The University of British Columbia Department of Electrical and Computer Engineering VancouverBCV6T 1Z4 Canada Carleton University Department of Systems and Computer Engineering OttawaONK1S 5B6 Canada Shenzhen University College of Computer Science and Software Engineering Shenzhen518060 China

Perceiving scene depth and 3D structure is one of the key tasks for Internet of Video Things (IoVT) devices to understand and interact with the environment. Self-supervised monocular depth estimation has demonstrated significant potential in leveraging large-scale unlabeled datasets to achieve competitive performance, thereby playing an increasingly important role in depth estimation. Despite recent methods providing additional supervisory signals through self-distillation strategies to improve depth estimation, an effective method for generating pseudo-depth labels suitable for addressing occlusion issues among elements far from the camera remains unexplored. To address this limitation, we propose a Patch-based Self-Distillation Learning Framework to exploit the potential of self-supervised monocular depth estimation in recovering fine-grained scene depth. In the proposed framework, elements far from the camera within the input image are enlarged by enlarging and cropping operations in the patch-based self-distillation branch. Guided by photometric consistency, the model learns the detailed occlusion relationships among elements from the enlarged patches, producing patch depth maps with fine structures. In the main branch, which takes full-scale images as input, patch depth maps serve as pseudo-depth labels through self-distillation loss to provide additional supervisory signals for regions where photometric consistency fails to offer effective supervision. This forces the depth estimation network to recover fine structures of elements far from the camera in full-scale input images. Regarding the architecture of the depth estimation network, we introduce a bin-center prediction. In this prediction, a global aggregator based on self-attention provides additional scene structure queries for adaptive scene depth discretization. Finally, to encourage the model to explore more general cues for depth inference beyond road plane cues, we propose a PatchMix data augmentati

关键词： Image enhancement

来源：评论

学校读者我要写书评

暂无评论

Archer: Adaptive Memory Compression with Page-Association-Rule Awareness for High-Speed Response of Mobile Devices 23

Archer: Adaptive Memory Compression with Page-Association-Ru...

引用

23rd USENIX Conference on File and Storage Technologies, FAST 2025

作者： Li, Changlong Zhu, Zongwei Wang, Chao Liu, Fangming Xu, Fei Sha, Edwin H.-M. Zhou, Xuehai School of Computer Science and Technology East China Normal University China Jianghuai Advance Technology Center Hefei230026 China MoE Engineering Research Center of Hardware/Software Co-Design Technology and Application China School of Software Engineering University of Science and Technology of China Hefei230026 China Suzhou Institute for Advanced Research University of Science and Technology of China Suzhou215123 China Huazhong University of Science and Technology China Peng Cheng Laboratory China

ISBN: (纸本)9781939133458

In mobile systems, memory can be compressed page-by-page to save space. This approach is widely adopted because memory data is accessed by page. However, this paper shows that the system response speed is significantly limited by page-grained compression. In this paper, we observe that approximately a quarter of anonymous memory pages are highly correlated, even though the association is implicit. Inspired by this, we propose Archer, an association-rule-aware memory compression framework in mobile systems. Archer demonstrates that memory in mobile devices should be compressed using flexible granularity, rather than relying solely on traditional page compression. To further integrate association-rule mining techniques into system design, we redesign the LRU mechanism and propose an adaptive memory compression region. Experimental results show that the average app launching speed is 1.55x faster when enabling Archer, and the average photographic speed and frame rate increase by 1.42x and 1.31x, respectively, compared to the state-of-the-art. © 2025 FAST. All Rights Reserved.

关键词： Portable equipment

来源：评论

学校读者我要写书评

暂无评论

Multi-scale convolutional fusion network embedded with improved attention mechanism for polarimetric synthetic aperture radar image classification

引用

Journal of Applied Remote Sensing 2025年第1期19卷

作者： Li, Jiewen Zhu, Yanan Ai, Jiaqiu Zhao, Zhicheng Shi, Jun Hefei University of Technology School of Computer Science and Information Engineering Hefei China Anhui University Information Materials and Intelligent Sensing Laboratory of Anhui Province Hefei China Hefei University of Technology School of Artificial Intelligence Hefei China Hefei University of Technology School of Software Hefei China

Convolutional neural networks (CNNs) have been widely applied in the field of polarimetric synthetic aperture radar (PolSAR) image classification and have achieved good classification results. However, due to the complexity of polarimetric information, traditional CNN structures are difficult to fully extract and integrate information from multiple polarimetric channels, resulting in insufficient feature diversity, which limits further improvement in classification accuracy. To this problem, this article proposes a multi-scale convolutional fusion network (MCF-Net). MCF-Net employs convolutional kernels of various sizes to perform parallel convolutions, extracting multi-scale features and optimally fusing them, thereby greatly enhancing the diversity and completeness of feature representation. In addition, to enhance the network's focus on important features and suppress irrelevant information, this article proposes an improved squeeze-and-excitation (SE) module, guided by mean, entropy, and Z-score (MEZ-SE), to weigh the importance of feature maps. By embedding the MEZ-SE module into MCF-Net, MEZ-MCF-Net significantly improves the completeness of feature representation and enhances the ability to distinguish different land cover categories. On the Flevoland, San Francisco, and Oberfaffenhofen PolSAR datasets, the overall accuracy of MEZ-MCF-Net exceeds that of the second-best method by 3.04%, 2.09%, and 1.99%, respectively. There is no doubt that MEZ-MCF-Net is an effective method for PolSAR image classification. © 2025 Society of Photo-Optical Instrumentation Engineers (SPIE).

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Role of Artificial Intelligence in Bioinformatics and Information Extraction Systems

Decision Making: Applications in Management and Engineering

引用

Decision Making: Applications in Management and engineering 2025年第1期8卷 401-418页

作者： Al-Safarini, Maram Y. Baashirah, Rania A. Computer Science Department Zarqa University Jordan Software Engineering Department University of Business & Technology Saudi Arabia

Artificial Intelligence (AI) is transforming numerous domains, including bioinformatics and information extraction systems, by advancing data processing capabilities, enhancing precision, and facilitating automation. The primary aim of this research is to examine the function of AI within the realms of bioinformatics and information extraction through a combination of quantitative and qualitative approaches. The quantitative component involved a sample of 152 participants, recognised as experts in bioinformatics and data science. The survey findings underscore the prevalent integration of AI in bioinformatics, particularly through the utilisation of reinforcement learning, neural networks, and natural language processing (NLP). Furthermore, AI substantially improves the analysis of biological data;however, it encounters challenges such as limited model interpretability and the lack of data standardisation. Notwithstanding these obstacles, AI-driven innovations in disease prognosis, personalised healthcare, and pharmaceutical development are anticipated to shape the future trajectory of bioinformatics. The qualitative findings, derived through thematic analysis encompassing core themes and sub-themes, reveal that AI significantly contributes to accelerating and refining information extraction processes. Technologies such as machine learning, NLP, and neural networks are instrumental in enhancing data processing efficiency. Nonetheless, issues such as inadequate data quality, elevated computational expenses, and the intricacy of AI models remain persistent. Looking ahead, AI is projected to integrate more effectively with big data infrastructures, enable real-time information extraction, and deliver increasingly tailored solutions. The study concludes with several policy recommendations to guide future implementation strategies. © 2025 Regional Association for Security and crisis management. All rights reserved.

关键词： Data integration

来源：评论

学校读者我要写书评

暂无评论

An Overview of Security Threats, Attack Detection and Defense for Large-Scale Multi-Agent Systems (LSMAS) in Internet of Things (IoT)

IEEE Transactions on Industrial Cyber-Physical Systems

引用

IEEE Transactions on Industrial Cyber-Physical Systems 2025年 3卷 70-81页

作者： Cai, Jijing Wen, Long Feng, Hailin Fang, Kai Chen, Junxin Wei, Wei Wang, Wei Zhejiang A and F University College of Mathematics and Computer Science Hangzhou311300 China Macau University of Science and Technology School of Innovation Engineering Macau999078 China Dalian University of Technology School of Software Dalian116621 China Xi'an University of Technology School of Computer Science and Engineering Xi'an710048 China Beijing Institute of Technology School of Medical Technology Beijing100081 China Artificial Intelligence Research Institute Shenzhen MSU-BIT University Guangdong-Hong Kong-Macao Joint Laboratory for Emotion Intelligence and Pervasive Computing Shenzhen518172 China

With the global advancement of the Internet of Things (IoT), large-scale multi-agent systems (LSMAS) technology has been increasingly adopted across various industries. Despite the widespread use of IoT, vulnerabilities in its software and hardware components pose significant challenges to ensuring security. To address the security concerns faced by LSMAS, this article explores its application in IoT domains such as smart grids, smart manufacturing, and smart healthcare. By examining the structural layers of IoT-the perception layer, network layer, and application layer-this paper analyzes the specific security threats encountered by each layer, along with the defense strategies designed to counter these attacks. Furthermore, we categorize and compare both traditional and AI-based attack detection methods, dividing the latter into machine learning-based, deep learning-based, and transfer learning-based approaches. This analysis highlights the strengths and weaknesses of the various attack detection methods currently deployed in LSMAS, identifying the existing challenges they face. Finally, we outline potential trends for the future development of LSMAS in IoT, offering research directions for experts in related fields. © 2023 IEEE.

关键词： Multi agent systems

来源：评论

学校读者我要写书评

暂无评论

Comparative Analysis of Differential Privacy Implementations on Synthetic Data 15

Comparative Analysis of Differential Privacy Implementations...

引用

15th IEEE Annual Computing and Communication Workshop and Conference, CCWC 2025

作者： Said, Huwida E. Mahmoud, Qusay H. Goyal, Mandeep Hashim, Faiza College of Technological Innovation Zayed University Department of Computing and Applied Technolgy Dubai United Arab Emirates Ontario Tech University Department of Electrical Computer and Software Engineering OshawaONL1G 0C6 Canada College of Information Technology Uae University Department of Computer Science and Software Engineering AI-Ain United Arab Emirates

ISBN: (纸本)9798331507695

Differential privacy offers a promising solution to balance data utility and user privacy. This paper compares two prominent differential privacy tools-PyDP and IBM's diffprivlib-that are applied to a synthetic dataset with medical attributes. We evaluate these tools based on their effectiveness in maintaining data privacy while preserving the statistical integrity of the data. Our results reveal that PyDP provides synthetic data that closely matches real-world data, making it appropriate for tasks requiring accuracy while striking a better balance between data utility and privacy. However, IBM's diffprivlib is more suitable for privacy-critical applications due to its stronger privacy guarantees, but at the cost of increased noise and decreased data utility. This paper contributes to the practical understanding of implementing differential privacy in machine learning and software applications and enhances the tools available for developers in sensitive data environments. © 2025 IEEE.

关键词： Data integrity

来源：评论

学校读者我要写书评

暂无评论

A comprehensive survey on 6G-security: physical connection and service layers

引用

Discover Internet of Things 2025年第1期5卷 1-30页

作者： Saeed, Mamoon M. Saeed, Rashid A. Hasan, Mohammad Kamrul Ali, Elmustafa Sayed Mazha, Tehseen Shahzad, Tariq Khan, Sunawar Hamam, Habib Sana00967 Yemen Department of Computer Engineering College of Computers and Information Technology Taif University P.O. Box 11099 Taif21944 Saudi Arabia Department of Electronics Engineering College of Electronics Engineering Sudan University of Science and Technology Omdurman Sudan Bangi43600 Malaysia Department of Electrical and Electronic Engineering Faculty of Engineering Red Sea University Port Sudan33312 Sudan School of Computer Science National College of Business Administration and Economics Lahore54000 Pakistan Faculty of Engineering Uni de Moncton MonctonNBE1A3E9 Canada Department of Computer Engineering COMSATS University Islamabad Sahiwal Campus Sahiwal57000 Pakistan Department of Computer Science and Software Engineering International Islamic University Islamabad44000 Pakistan School of Electrical Engineering University of Johannesburg Johannesburg2006 South Africa BP 1989 Av. Grandes Ecoles Libreville Gabon Bridges for Academic Excellence-Spectrum Tunis Tunisia Department of Computer Science School Education Department Government of Punjab Layyah31200 Pakistan Chitkara University Institute of Engineering and Technology Chitkara University Punjab Rajpura140401 India

After adopting 5G technology, businesses and academia have started working on sixth-generation wireless networking (6G) technologies. Mobile communications options are expected to expand in areas where previous generations could not do so. 6G networks are anticipated to be constructed using various diverse technologies. These encompass diverse cutting-edge advancements, such as distributed ledger systems like blockchain, visible light communications (VLC), post-quantum cryptography, edge computing, molecular communication, THz, and other advances. These advances necessitate a reassessment of previous security strategies from a security perspective. In the future, networks must adhere to stricter criteria for authentication, encryption, access control, connectivity, and detection of harmful activities. Ensuring privacy and dependability necessitates the implementation of supplementary security protocols. The essay explores the primary concerns and challenges related to the security of the 6G network. This paper describes the improvements in security in communications from 1G through 6G. This paper divides security in the sixth generation into three layers: physical, connection, and service. Each layer-by-layer discusses the standard technologies and security issues for each technology proposed in each sixth-generation security layer. All proposed solutions for each of the three layers are discussed in Sixth Generation Security. It also reviews all proposed solutions for each layer, indicating the proposed solution and its limitations. © The Author(s) 2025.

关键词： 5G mobile communication systems

来源：评论

学校读者我要写书评

暂无评论

Trauma-Informed Insights from Co-Design of Self-Disclosure Robots with Domestic Abuse Survivors 20

Trauma-Informed Insights from Co-Design of Self-Disclosure R...

引用

20th Annual ACM/IEEE International Conference on Human-Robot Interaction, HRI 2025

作者： Morris, Nyomi Carrine Williams, Tom Jelen, Ben Colorado School of Mines MIRRORLab GoldenCO United States Rose-Hulman Institute of Technology Department of Computer Science and Software Engineering Terre HauteIN United States

ISBN: (纸本)9798350378931

Survivors of domestic abuse face significant challenges securing recovery resources, such as housing, mental health care, and social connections. Accessing these resources requires survivors to disclose their status as survivors of domestic abuse;a process that can be traumatic and emotionally burdensome. In this work, we consider how social robots might help domestic abuse survivors to address challenges surrounding self-disclosure. To do so, we conducted a three-session co-design workshop involving 8 participants from a domestic abuse shelter. Our results provide three key contributions: (1) new insights into the benefits and barriers to self-disclosure;(2) new insights into the ways that social robots can help survivors to better achieve those benefits and overcome those barriers before, during, and after disclosure, including key design recommendations associated with each of these phases;and (3) observations into how a trauma-informed computing perspective may enable more effective design work in Human-Robot Interaction. © 2025 IEEE.

关键词： Social robots

来源：评论

学校读者我要写书评

暂无评论

COSIS: An AI-Enabled Digital Transformation Framework Integrating Large Language Models and Key Performance Indicators 21st

COSIS: An AI-Enabled Digital Transformation Framework Integr...

引用

21st International Conference on Services Computing, SCC 2024, Held as Part of the Services Conference Federation, SCF 2024

作者： Zhang, Liang-Jie Chen, Huan He, Sheng Li, Changhu Chen, Junyang Zhang, Haodi Du, Wenfeng Center for AI Services Computing College of Computer Science and Software Engineering Shenzhen University Shenzhen China SF Technology Co. Ltd. Shenzhen China Kingdee Research Kingdee International Software Group Co. Ltd. Shenzhen China

ISBN: (纸本)9783031769993

The rapid advancement of Artificial Intelligence (AI) and Large Language Models (LLMs) has significantly transformed various sectors of society, compelling enterprises to undertake comprehensive digital transformations to remain competitive in the AI era. This paper introduces COSIS, an AI-enabled digital trans-formation framework comprising five key dimensions: Culture, Operation, Strategy, Innovation, and Service. The COSIS framework provides a structured approach for enterprises to navigate the complexities of digital transformation by integrating AI and LLM technologies into their core functions. For each dimension, the paper identifies specific key performance indicators (KPIs) to measure the effectiveness of AI-enabled transformations. These KPIs enable enterprises to quantitatively assess progress and outcomes, ensuring alignment with strategic objectives and facilitating continuous improvement. Additionally, the paper discusses the implications of AI and LLM technologies on enterprise management team building in the AI era. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Human resource management

来源：评论

学校读者我要写书评

暂无评论

Batch Execution of Microbenchmarks for Efficient Performance Testing

Batch Execution of Microbenchmarks for Efficient Performance...

引用

International Conference on software Testing, Verification, and Validation, ICST

作者： Mostafa Jangali Kundi Yao Yiming Tang Diego Elias Costa Weiyi Shang Department of Electrical and Computer Engineering University of Waterloo Waterloo Canada Department of Software Engineering Rochester Institute of Technology USA Computer Science and Software Engineering Concordia University Montreal Canada

ISBN: (数字)9798331508142

ISBN: (纸本)9798331508159

Performance microbenchmarking is essential for ensuring software quality by providing granular insights into code efficiency. While automated performance microbenchmark generation tools (e.g., ju2jmh) are proposed to alleviate practitioners from manually curating microbenchmarks, the high volume of generated benchmarks can lead to protracted benchmarking execution time, as many of the generated benchmarks are too short in nature to be valuable for evaluating performance. In this paper, we present a novel approach that optimizes microbenchmark execution through a batching strategy, i.e., grouping benchmarks with similar code coverage and treating them as a single unit to 1) reduce execution overhead and 2) reduce the bias from microbenchmarks that are too short. We evaluate the effectiveness of this enhancement across various Java projects, comparing the execution times of clustered and individual micro benchmarks. Our findings demonstrate substantial improvements in execution efficiency, reducing execution time by up to 89.81% while preserving high microbenchmark stability.

关键词： software testing Java Codes software quality software performance Benchmark testing

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：