检索结果-内蒙古大学图书馆

Black-box domain adaptation for cross-domain on-device machinery fault diagnosis via hierarchical debiased self-supervised learning

学校读者我要写书评

暂无评论

Knowledge-Based Systems 2025年 324卷

作者： Mengliang Zhu Jie Liu Yanglong Lu Zhongxu Hu Li Yuan Kaibo Zhou MOE Key Laboratory of Image Information Processing and Intelligent Control Belt and Road Joint Laboratory on Measurement and Control Technology School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan 430074 China National Center of Technology Innovation for Digital Construction School of Civil and Hydraulic Engineering Huazhong University of Science and Technology Wuhan 430074 China Department of Mechanical and Aerospace Engineering The Hong Kong University of Science and Technology Clear Water Bay Hong Kong SAR China School of Mechanical Science and Engineering Huazhong University of Science and Technology Wuhan 430074 China

This paper explores black-box domain adaptation (BBDA) for cross-domain on-device machinery fault diagnosis. Specifically, a pre-trained black-box source model is deployed on a cloud platform with only its input-output API accessible, and a randomly initialized target model is trained locally utilizing unlabeled target data on an edge device. In BBDA, neither raw data nor model parameters are transmitted across domains, and the only available supervised information for adaptation is API-queried predictions on unlabeled target data. However, API-queried predictions are inevitably noisy and class-imbalanced due to cross-domain distribution discrepancy and inherent category bias. Directly utilizing them for on-device training easily leads to confirmation bias and exacerbates category bias, thereby severely degrading performance. This paper proposes a hierarchical debiased self-supervised learning framework for BBDA from a novel research perspective of concurrently alleviating both confirmation and category biases. Specifically, debiased knowledge distillation is proposed to gradually enhance discriminability of the randomly initialized target model by providing debiased training predictions. An adaptive category-unbiased sample division strategy is devised to divide unlabeled target data into certain-aware and uncertain-aware sets, enabling fully exploiting intrinsic data structures and utilizing underlying sample relationships within the target domain. On this basis, debiased prototypical self-training is proposed for the certain-aware set to obtain reliable pseudo-labels for self-training, leveraging global semantic structure. Hierarchical neighborhood adaptation is customized for the uncertain-aware set to enforce the cluster assumption for noise refinement, exploiting local clustering structure and utilizing underlying sample relationships. Extensive experiments and analyses on public and practical datasets validate the proposed framework’s superiority in cross-dom

关键词： Black-box domain adaptation Category bias Confirmation bias Cross-domain on-device machinery fault diagnosis Self-supervised learning

A polynomial chaos approach to robust H∞ static output-feedback control with bounded truncation error

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Wan, Yiming Shen, Dongying E. Lucia, Sergio Findeisen, Rolf Braatz, Richard D. Artificial Intelligence and Automation Huazhong University of Science and Technology Key Laboratory of Image Processing and Intelligent Control Ministry of Education Wuhan430074 China Massachusetts Institute of Technology 77 Massachusetts Avenue CambridgeMA02139 United States TU Dortmund University Dortmund44227 Germany Otto-von-Guericke University Magdeburg Magdeburg39106 Germany

This article considers the H∞ static output-feedback control for linear time-invariant uncertain systems with polynomial dependence on probabilistic time-invariant parametric uncertainties. By applying polynomial chaos theory, the control synthesis problem is solved using a high-dimensional expanded system which characterizes stochastic state uncertainty propagation. A closed-loop polynomial chaos transformation is proposed to derive the closed-loop expanded system. The approach explicitly accounts for the closed-loop dynamics and preserves the L2-induced gain, which results in smaller transformation errors compared to existing polynomial chaos transformations. The effect of using finite-degree polynomial chaos expansions is first captured by a norm-bounded linear differential inclusion, and then addressed by formulating a robust polynomial chaos based control synthesis problem. This proposed approach avoids the use of high-degree polynomial chaos expansions to alleviate the destabilizing effect of truncation errors, which significantly reduces computational complexity. In addition, some analysis is given for the condition under which the robustly stabilized expanded system implies the robust stability of the original system. A numerical example illustrates the effectiveness of the proposed approach. © 2021, CC BY.

关键词： Feedback control

Hand Gesture Recognition Based on Multi-Classification Adaptive Neuro-Fuzzy Inference System and pMMG

学校读者我要写书评

暂无评论

Hand Gesture Recognition Based on Multi-Classification Adapt...

International Conference on Advanced Robotics and Mechatronics (ICARM)

作者： Lei Wang Jian Huang Dongrui Wu Tao Duan Rui Zong Shicong Jiang Key Laboratory of Image Processing and Intelligent Control School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan China Laboratory for Economics and Computation City University of Hong Kong Kowloon Hong Kong SAR China

ISBN: (数字)9781728164793

ISBN: (纸本)9781728164809

In this paper, a multi-classification adaptive neuro-fuzzy inference system combining neural-network and a TSK fuzzy system is proposed to recognize six commonly used gestures. Several techniques including mini-batch gradient descent with L2 regularization, DropRule and AdaBound are integrated to improve the generalization ability of the system and the efficiency of training. Numerical results show that the average classification accuracy of the multiple classifier systems is 95.12%, and this value is higher than some other multiple classifier systems (CNN, LDA, etc.) using MMG signals as the inputs for hand gesture recognition.

关键词： Training Electromyography Muscles Skin Robots Mechatronics Adaptive systems

Resilient Distributed Predefined Time Secondary control for Cyber-Physical Microgrids

学校读者我要写书评

暂无评论

IET Renewable Power Generation 2025年第1期19卷

作者： Junfeng Tan Fan Zhang Yanlu Huang Shuai Zhao Hongyu Su China Southern Power Grid Digital Grid Research Institute Co. Ltd. China Southern Power Grid Artificial Intelligence Technology Co. Ltd. Guangzhou China School of Artificial Intelligence and Automation and Technology and also with the Key Laboratory of Image Processing and Intelligent Control Ministry of Education Huazhong University of Science and Technology Wuhan China

This paper proposed a resilient distributed predefined-time sliding mode control for islanded AC microgrids with external disturbances caused by noisy circumstances or cyber-attacks. By utilizing the predefined-time convergence theory, the voltage regulation and frequency restoration as well as active power sharing can be achieved within a predefined time, which is directly equal to an adjustable parameter. Furthermore, based on the integral sliding mode control approach, the proposed method can completely compensate the external disturbance. Different from some voltage control methods based on complex second-order consensus, a novel secondary controller is designed by adopting the virtual control technique, such that the voltage regulation can be achieved under a first-order consensus with a corresponding tracking controller. In addition, the direct Lyapunov method is utilized to prove the stability of islanded AC microgrids under the proposed controller, and the analysis of predefined-time convergence is also given. Finally, case studies on a microgrid test system with four distributed generator is built in the MATLAB/SimPowerSystems software environment are conducted to demonstrate the effectiveness and superior performance of the proposed control scheme.

关键词： distributed power generation Microgrids robust control

Pool-Based Unsupervised Active Learning for Regression Using Iterative Representativeness-Diversity Maximization (iRDM)

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Liu, Ziang Jiang, Xue Luo, Hanbin Fang, Weili Liu, Jiajing Wu, Dongrui Key Laboratory of the Ministry of Education for Image Processing and Intelligent Control School of Artificial Intelligence and Automation Huazhong University of Science and Technology China School of Civil Engineering and Mechanics Huazhong University of Science and Technology China

Active learning (AL) selects the most beneficial unlabeled samples to label, and hence a better machine learning model can be trained from the same number of labeled samples. Most existing active learning for regression (ALR) approaches are supervised, which means the sampling process must use some label information, or an existing regression model. This paper considers completely unsupervised ALR, i.e., how to select the samples to label without knowing any true label information. We propose a novel unsupervised ALR approach, iterative representativeness-diversity maximization (iRDM), to optimally balance the representativeness and the diversity of the selected samples. Experiments on 12 datasets from various domains demonstrated its effectiveness. Our iRDM can be applied to both linear regression and kernel regression, and it even significantly outperforms supervised ALR when the number of labeled samples is small. Copyright © 2020, The Authors. All rights reserved.

关键词： Regression analysis

Transfer Learning for Motor imagery Based Brain-Computer Interfaces: A Complete Pipeline

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Wu, Dongrui Jiang, Xue Peng, Ruimin Kong, Wanzeng Huang, Jian Zeng, Zhigang Key Laboratory of the Ministry of Education for Image Processing and Intelligent Control School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China Zhejiang Key Laboratory for Brain-Machine Collaborative Intelligence Hangzhou Dianzi University Hangzhou310018 China

Transfer learning (TL) has been widely used in motor imagery (MI) based brain-computer interfaces (BCIs) to reduce the calibration effort for a new subject, and demonstrated promising performance. While a closed-loop MI-based BCI system, after electroencephalogram (EEG) signal acquisition and temporal filtering, includes spatial filtering, feature engineering, and classification blocks before sending out the control signal to an external device, previous approaches only considered TL in one or two such components. This paper proposes that TL could be considered in all three components (spatial filtering, feature engineering, and classification) of MI-based BCIs. Furthermore, it is also very important to specifically add a data alignment component before spatial filtering to make the data from different subjects more consistent, and hence to facilitate subsequential TL. Offline calibration experiments on two MI datasets verified our proposal. Especially, integrating data alignment and sophisticated TL approaches can significantly improve the classification performance, and hence greatly reduces the calibration effort. Copyright © 2020, The Authors. All rights reserved.

关键词： Electroencephalography

Closer to Pre-trained Network Transfer Better

学校读者我要写书评

暂无评论

Closer to Pre-trained Network Transfer Better

IEEE Joint International Information Technology and Artificial Intelligence Conference (ITAIC)

作者： Siyu Chen Wei Li School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan People’s Republic of China Image Processing and Intelligent Control Key Laboratory Education Ministry of China Wuhan People’s Republic of China

ISBN: (数字)9781728152448

ISBN: (纸本)9781728152455

In recent years, Deep Neural Network (DNN) has been widely used in the domain of computer vision, but its further development is restricted because of the lack of train samples. Fine-tuning is one of deep transfer learning methods to solve this problem. Nowadays, some existing Fine-tuning approaches assume that if the fine-tuned network could be as close to the pre-trained network as possible, the source knowledge would be fully utilized to improve the study effect. Although these approaches work pretty well, the assumptions on which they are based have not been systematically proved. Due to the complexity of DNNs, offering a reasonable and accurate explanation of the performance of the network is significant for designing new algorithms. In this paper, we design a series of experiments with or without transfer learning respectively, to characterize the relationship between network differences and network performance. The results show that there is an obvious negative correlation between them under certain circumstances. This conclusion can not only provide a more reasonable explanation for the relevant transfer learning algorithm, but also provide support and inspiration for the future work.

关键词： Knowledge engineering Correlation Transfer learning Neural networks Task analysis Information technology Knowledge transfer

NiuEM: A Nested-iterative Unsupervised Learning Model for Single-particle Cryo-EM image processing

学校读者我要写书评

暂无评论

NiuEM: A Nested-iterative Unsupervised Learning Model for Si...

IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

作者： Rui Hu Jiaming Cai Wangjie Zheng Yang Yang Hong-Bin Shen Shanghai Jiao Tong University and Key Laboratory of Shanghai Education Commission for Intelligent Interaction and Cognitive Engineering Shanghai China Institute of Image Processing and Pattern Recognition Shanghai Jiao Tong University and Key Laboratory of System Control and Information Processing Ministry of Education of China Shanghai China Shanghai Jiao Tong University Shanghai China

ISBN: (数字)9781728162157

ISBN: (纸本)9781728162164

Cryo-electron microscopy (cryo-EM) has become a mainstream technology for solving spatial structures of biomacromolecules, while the processing of cryo-EM images is a very challenging task. One of the great challenges is the high noise in the images. A common method is to cluster the images with close projecting angles to get mean images, which are used for 3D reconstruction. However, due to the extremely low signal-to-noise-ratio, common clustering methods often fail to obtain high-quality mean images, leading to poorly reconstructed structures. In this study, we present a new unsupervised learning framework, called NiuEM, to discriminate images captured from different angles and yield cluster-mean images. NiuEM first generates pseudo-labels and then exploits both contrastive loss and cross-entropy loss for training convolutional layers to learn feature representations. Moreover, the pseudo-labels are updated iteratively to enhance the reliability of labels. We assess the performance of NiuEM on four data sets via both visualized and quantitative experiments. Especially, two kinds of metrics are adopted to measure the performance, regarding the clustering quality and the resolution of reconstructed 3D models, respectively. The experimental results show that NiuEM achieves very competitive clustering accuracy in the comparison with the state-of-the-art image clustering methods. Moreover, the cluster mean images yielded by NiuEM lead to better initial 3D models compared with the mainstream reconstruction tools.

关键词： Three-dimensional displays Feature extraction image reconstruction Task analysis Solid modeling Electron microscopy Convolutional neural networks

A survey on negative transfer

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Zhang, Wen Deng, Lingfei Zhang, Lei Wu, Dongrui the Key Laboratory of the Ministry of Education for Image Processing and Intelligent Control School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China the School of Microelectronics and Communication Engineering Chongqing University Chongqing400044 China

—Transfer learning (TL) utilizes data or knowledge from one or more source domains to facilitate the learning in a target domain. It is particularly useful when the target domain has very few or no labeled data, due to annotation expense, privacy concerns, etc. Unfortunately, the effectiveness of TL is not always guaranteed. Negative transfer (NT), i.e., leveraging source domain data/knowledge undesirably reduces the learning performance in the target domain, has been a long-standing and challenging problem in TL. Various approaches have been proposed in the literature to handle it. However, there does not exist a systematic survey on the formulation of NT, the factors leading to NT, and the algorithms that mitigate NT. This paper fills this gap, by first introducing the definition of NT and its factors, then reviewing about fifty representative approaches for overcoming NT, according to four categories: secure transfer, domain similarity estimation, distant transfer, and NT mitigation. NT in related fields, e.g., multi-task learning, lifelong learning, and adversarial attacks, are also discussed. Copyright © 2020, The Authors. All rights reserved.

关键词： Surveys