检索结果-内蒙古大学图书馆

41st International Conference on Machine Learning, ICML 2024

作者： Liu, Shikun Zou, Deyu Zhao, Han Li, Pan Department of Electrical and Computer Engineering Georgia Institute of Technology GA United States School of Data Science University of Science and Technology of China Hefei China Department of Computer Science University of Illinois Urbana-Champaign Champaign United States

Graph-based methods, pivotal for label inference over interconnected objects in many real-world applications, often encounter generalization challenges, if the graph used for model training differs significantly from the graph used for testing. This work delves into Graph Domain Adaptation (GDA) to address the unique complexities of distribution shifts over graph data, where interconnected data points experience shifts in features, labels, and in particular, connecting patterns. We propose a novel, theoretically principled method, Pairwise Alignment (Pair-Align) to counter graph structure shift by mitigating conditional structure shift (CSS) and label shift (LS). Pair-Align uses edge weights to recalibrate the influence among neighboring nodes to handle CSS and adjusts the classification loss with label weights to handle LS. Our method demonstrates superior performance in real-world applications, including node classification with region shift in social networks, and the pileup mitigation task in particle colliding experiments. For the first application, we also curate the largest dataset by far for GDA studies. Our method shows strong performance in synthetic and other existing benchmark datasets. Copyright 2024 by the author(s)

关键词： Graph theory

来源：评论

学校读者我要写书评

暂无评论

Robust Tracking of Dynamic Visual Object with Complex Environmental Factors 24

Robust Tracking of Dynamic Visual Object with Complex Enviro...

引用

2024 7th International Conference on computer Information science and Artificial Intelligence, CISAI 2024

作者： Zhang, Jiahe Wang, Bingxiang Niu, Jinyuan School of Computer Science and Technology Xi'an Jiaotong University Shaanxi Xi'an China School of Mechanical Engineering Xi'an Jiaotong University Shaanxi Xi'an China

ISBN: (纸本)9798400707254

Aiming at a series of problems caused by complex environment changes, such as target tracking instability, target retracking deviation after occluding or disappearing and high similarity interference, a target tracking algorithm based on TLD framework is proposed in this paper. By using CLAHE(Contrast Limited Adaptive Histogram Equalization) image enhancement technology to preprocess the image, the color contrast of different pixels can be enlarged to increase the image recognition ability. To ensure the accuracy and real-time performance of the model, Kalman filter is used to predict the possible position of the target when it is blocked or disappeared, and detect and track the target immediately after it reappears. Then, the Median-Flow optical flow tracker with CSRT assisted TLD algorithm makes use of its dual reliability mechanism of channel and space. As the result, the target tracking task can be fulfilled effectively and stably. Experimental results show that the proposed algorithm can achieve high tracking accuracy and robustness under complex environment. © 2024 Copyright held by the owner/author(s).

关键词： Kalman filters

来源：评论

学校读者我要写书评

暂无评论

Conceptualizing Processes in Robotic Process Automation: A Basic Meta-Model 10

Conceptualizing Processes in Robotic Process Automation: A B...

引用

10th International Conference on Systems and Informatics, ICSAI 2024

作者： Zhang, Lingtong Yang, Jing Ouyang, Chun Yu, Yang Pan, Maolin School of Computer Science and Engineering Sun Yat-sen University GuangZhou China Queensland University of Technology Australia School of Software Engineering Sun Yat-sen University ZhuHai China

ISBN: (纸本)9798331510138

Robotic Process Automation (RPA) uses software to mimic human actions in performing repetitive, time-consuming tasks, enabling organizations to strategically reallocate their workforce towards more value-driven activities. RPA can be used to build new intelligent systems that combine business with Artificial Intelligence (AI). However, despite its widespread adoption, RPA lacks standardization and a theoretical foundation, giving rise to misconception. For example, it has caused confusion between RPA and the established field of Business Process Automation (BPA), which has prevented organizations from properly using RPA in new intelligent systems. In this paper, we address the gap by proposing a basic meta-model for conceptualizing processes in RPA. Our research involves identifying key characteristics of processes suitable for RPA, deriving fundamental concepts of processes in RPA, and articulating the interrelations between them. Built upon such a conceptual foundation, we conduct a comparative analysis between RPA and BPA, clarifying the differences between the two. We further demonstrate the practical applicability of our proposed basic meta-model by identifying an instantiation of the meta-model manifested in three mainstream RPA software products, respectively. Our research contributes to laying the conceptual foundations of RPA and facilitating its standardization. © 2024 IEEE.

关键词： Intelligent robots

来源：评论

学校读者我要写书评

暂无评论

Fast Algorithm Based on Non-first Pre-coding Skip for AVS3 Intra Coding 8

Fast Algorithm Based on Non-first Pre-coding Skip for AVS3 I...

引用

8th International Conference on Signal and Image Processing, ICSIP 2023

作者： Cao, Xueyan Lin, Tao Zhao, Liping Yang, Yufen Zhou, Kailun Wang, Shuhui Tongji University College of Electronic and Information Engineering Shanghai China Shaoxing University School of Computer Science and Engineering Shaoxing China Shanghai Institute of Technology School of Computer Science and Information Engineering Shanghai China

ISBN: (纸本)9798350397932

The third generation of Audio Video coding Standard (AVS3) adopts multi-type tree, including Quad-Tree (QT), Binary-Tree (BT) plus Extend Quad-Tree (EQT), partition structure. During multi-type tree recursive pre-coding process of encoding, each Coding Units (CU) of each Size at each Location (CUeSeL) of the picture is repeatedly pre-coded many times when visiting each branch of each tree. Although there is early termination mechanism, the number of visit for each CUeSeL is still very high and up to more than twenty. This process brings very high coding complexity. To reduce the coding complexity of AVS3, this paper proposes a method named non-first pre-coding skip (NFPS) for CUeSeL. The contributions of the paper are as follows. In I picture, for a CU with top-left location (x, y), if both x and y are multiples of 64, then each non-first pre-coding is skipped and the coding result of the first pre-coding is reused for each non-first pre-coding. In non-I picture, for a CU with top-left location (x, y), if both x and y are multiples of 64 and a set of conditions is satisfied, then each non-first pre-coding is skipped and the coding result of the first pre-coding is reused for each non-first pre-coding. The experimental results demonstrate that NFPS has a negligible impact on coding efficiency under All intra (AI), Low Delay B (LDB), Random Access (RA) configurations, and coding runtime is reduced by 6% for AI and almost not changed for LDB and RA. © 2023 IEEE.

关键词： Location

来源：评论

学校读者我要写书评

暂无评论

Real-Time Lightweight Convolutional Neural Network for Polyp Detection in Endoscope Images

引用

Journal of Shanghai Jiaotong university (science) 2025年第3期30卷 521-534页

作者： Si, Bingqi Pang, Chenxi Wang, Zhiwu Jiang, Pingping Yan, Guozheng School of Electronic Information and Electrical Engineering Shanghai Jiao Tong University Shanghai200240 China Shanghai Engineering Research Center of Intelligent Drug Detoxification and Rehabilitation Shanghai200240 China College of Computer Science and Technology Jilin University Changchun130012 China

Colorectal cancer is the most common cancer with a second mortality rate. Polyp lesion is a precursor symptom of colorectal cancer. Detection and removal of polyps can effectively reduce the mortality of patients in the early period. However, mass images will be generated during an endoscopy, which will greatly increase the workload of doctors, and long-term mechanical screening of endoscopy images will also lead to a high misdiagnosis rate. Aiming at the problem that computer-aided diagnosis models deeply depend on the computational power in the polyp detection task, we propose a lightweight model, coordinate attention-YOLOv5-Lite-Prune, based on the YOLOv5 algorithm, which is different from state-of-the-art methods proposed by the existing research that applied object detection models or their variants directly to prediction task without any lightweight processing, such as faster region-based convolutional neural networks, YOLOv3, YOLOv4, and single shot multibox detector. The innovations of our model are as follows: First, the lightweight EfficientNetLite network is introduced as the new feature extraction network. Second, the depthwise separable convolution and its improved modules with different attention mechanisms are used to replace the standard convolution in the detection head structure. Then, the α-intersection over union loss function is applied to improve the precision and convergence speed of the model. Finally, the model size is compressed with a pruning algorithm. Our model effectively reduces parameter amount and computational complexity without significant accuracy loss. Therefore, the model can be successfully deployed on the embedded deep learning platform, and detect polyps with a speed above 30 frames per second, which means the model gets rid of the limitation that deep learning models must rely on high-performance servers. © Shanghai Jiao Tong university 2023.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Protecting machine learning systems using blockchain: solutions, challenges and future prospects

引用

Multimedia Tools and Applications 2024年 1-28页

作者： Hajlaoui, Rejab Dhahri, Salah Mahfoudhi, Sami Moulahi, Tarek Alotibi, Gaseb Department of Computer and Informtion Sciences College of Computer Science and Engineering University of Hail Hail City Saudi Arabia Laboratory of Advanced Multifunctional Materials and Technological Applications Faculty of Science and Technology of Sidi Bouzid University of Kairouan Sidi Bouzid Tunisia Department of Information and Technology Management Mediterranean School of Business South Mediterranean University Tunis Tunisia Department of Information Technology College of Computer Qassim University Buraidah Saudi Arabia Department of Computers and Information Technology Alwajah College University of Tabuk Tabuk Saudi Arabia

Machine learning-based systems have emerged as the primary means for achieving the highest levels of productivity and efficiency. They have become the most influential competitive factor for many technologies and business companies such as Cloud AI Companies (Google, Facebook…) and Health Care AI Companies (Tempus, Nanox …). However, privacy and security issues have become the biggest challenge facing all ML applications. These threats are known as Adversarial Machine Learning (AML). Their main goal is to maliciously manipulate training data and model parameters to obtain misleading results. To overcome these challenges, several research studies have proven that Blockchain, which relies on cryptographic technologies, constitutes a promising solution for securing ML applications. It provides high guarantees to make the system scalable, reliable and more secure. In this survey, we provide an overview of ML applications and the most hostile attacks they face. Then, we define the blockchain technique and its distinctive features. After that, we conduct a comprehensive and detailed survey of the latest and best blockchain-based research to address security issues in ML applications (integrity, confidentiality and availability). We analyze, compare and comment on all the proposed solutions. In the end, we discuss in detail the various challenges that hinder the adoption of blockchain technology and we extract the most important scientific trends that will serve as guidance and support for researchers in their future works. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Blockchain

来源：评论

学校读者我要写书评

暂无评论

Machine Learning-Based Fault-Tolerant Inertial Navigation System for Micro Tunnel Boring Machines

Machine Learning-Based Fault-Tolerant Inertial Navigation Sy...

引用

2024 International Conference on Decision Aid sciences and Applications, DASA 2024

作者： Roudra, Nafis Ashraf Saha, Neelavro Sakib, Saadman Al Mamun, Rahat Shithil, Shaekh Mohammad Iqbal, M Shahriar Tasnim, Rumana BRAC University Department of Computer Science and Engineering Bangladesh Independent University Department of Computer Science and Engineering Bangladesh School of Mechanical Engineering Universiti Teknologi Malaysia Malaysia Rajshahi University of Engineering & Technology Department of Mechanical Engineering Bangladesh World University of Bangladesh Department of Mechatronics Engineering Bangladesh

ISBN: (纸本)9798350369106

Underground tunneling poses a navigation challenge as Micro Tunnel Boring Machines (mTBMs) operate in GPS denied environments. In such settings, Inertial Measurement Units (IMUs) offer an alternative navigation solution. However, faults in mTBM operation caused by various environmental and operational factors lead to inaccuracies in IMU data. This paper presents a machine learning-based fault tolerance system for IMU based mTBM navigation. IMU data corresponding to four specific mTBM faults: vibration, high temperature, magnetic interference and mTBM precession were collected under rigorous testing in a lab environment. Support Vector Machine (SVM) was applied to classify these faults and predict the mTBM's operational state based on the IMU readings. The results demonstrate the effectiveness of the SVM model in identifying and classifying the faults, thereby enhancing the reliability and safety of mTBM operations in challenging environments. © 2024 IEEE.

关键词： Inertial navigation systems

来源：评论

学校读者我要写书评

暂无评论

Traceable ring signature schemes based on SM2 digital signature algorithm and its applications in the data sharing scheme

引用

Frontiers of computer science 2024年第2期18卷 251-253页

作者： Yongxin ZHANG Hong LEI Bin WANG Qinghao WANG Ning LU Wenbo SHI Bangdao CHEN Qiuling YUE R&D Department SSC Holding CompanyLtd.Chengmai 571924China School of Cyberspace Security(School of Cryptography) Hainan UniversityHaikou 570228China College of Computer Science and Engineering Northeastern UniversityShenyang 110819China School of Computer Science and Technology Xidian UniversityXi'an 710071China Blockchain Research Group Oxford-Hainan Blockchain Research InstituteChengmai 571924China

1 Introduction Identity privacy concerns hinder data sharing by casting doubt on the safeguarding of personal information,eroding trust,and impeding the willingness of individuals and organizations to exchange their data[1,2].The traceable ring signatures(TRSs)addresses the contradiction between identity privacy and regulation[3],no scheme has been developed thus far that is based on SM2,the Chinese cryptographic public key algorithm standard,without relying on centralized trust.

关键词： SM2 signature scheme

来源：评论

学校读者我要写书评

暂无评论

Towards Adversarially Robust Continual Learning 48

Towards Adversarially Robust Continual Learning

引用

48th IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2023

作者： Bai, Tao Chen, Chen Lyu, Lingjuan Zhao, Jun Wen, Bihan Nanyang Technological University School of Computer Science and Engineering Singapore Zhejiang University College of Computer Science and Technology China Sony Ai Nanyang Technological University School of Electrical and Electronic Engineering Singapore

ISBN: (纸本)9781728163277

Recent studies show that models trained by continual learning can achieve the comparable performances as the standard supervised learning and the learning flexibility of continual learning models enables their wide applications in the real world. Deep learning models, however, are shown to be vulnerable to adversarial attacks. Though there are many studies on the model robustness in the context of standard supervised learning, protecting continual learning from adversarial attacks has not yet been investigated. To fill in this research gap, we are the first to study adversarial robustness in continual learning and propose a novel method called Task-Aware Boundary Augmentation (TABA) to boost the robustness of continual learning models. With extensive experiments on CIFAR-10 and CIFAR-100, we show the efficacy of adversarial training and TABA in defending adversarial attacks. © 2023 IEEE.

关键词： Adversarial training continual learning data augmentation

来源：评论

学校读者我要写书评

暂无评论

Improving Alzheimer's Disease Diagnosis on Brain MRI Scans with an Ensemble of Deep Learning Models 1

Improving Alzheimer's Disease Diagnosis on Brain MRI Scans w...

引用

1st International Conference on Artificial Intelligence for Innovations in Healthcare Industries, ICAIIHI 2023

作者： Reza, Md. Sazid Jahangir Kabir, Mir Md. Rakib Mollah, Md. Abdur Rajshahi University of Engineering and Technology Department of Computer Science and Engineering Rajshahi Bangladesh Td School University of Technology Sydney Data Science and Innovation Sydney Australia

ISBN: (纸本)9798350330915

Alzheimer's disease (AD) is a widespread neurolog-ical condition affecting millions globally. It gradually advances, leading to memory loss, cognitive deterioration, and a substantial decline in overall quality of life for those affected. AD patients experience memory decline, eroding cherished memories and straining relationships, while daily tasks become challenging. Numerous investigations have been conducted in this field, as the timely identification of Alzheimer's disease at its initial stage is of the utmost importance. A major limitation in this field is the predominant emphasis on using single fine-tuned CNN architecture or comparing pre-trained and custom CNN models for Alzheimer's detection, often on small datasets, which neglects a more comprehensive approach. Using smaller datasets can negatively impact deep learning modeling accuracy due to overfitting, limited representation, and poor generalization. This study addresses the current research problems and proposes an ensemble approach that combines predictions from various pre-trained models, including DenseNet-121, EfficientNet B7, ResNet-50, VGG-19, and also from a Custom CNN. The model averaging ensemble method was applied, a subset of the Stacking Ensemble, to two ADNI datasets, with Dataset-I being the larger. The goal was to assess the efficacy of this ensemble approach for accurate multiclass classification on ADNI datasets, where it successfully identified all classes despite differing sample volumes. A vast experiment was conducted on two distinct and widely recognized real-world datasets, resulting in accuracies of 99.96% and 98.90% respectively. Finally, the outcome of the research compared with recent research findings demonstrates the potential of our approach in advancing Alzheimer's disease detection by outperforming other benchmark approaches by a significant margin. © 2023 IEEE.

关键词： ADNI DenseNet-121 EfficientNet B7 Ensemble Model Averaging Predictions ResNet-50 VGG-19

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：