The success of deep transfer learning in fault diagnosis is attributed to the collection of high-quality labeled data from the source domain.However,in engineering scenarios,achieving such high-quality label annotatio...The success of deep transfer learning in fault diagnosis is attributed to the collection of high-quality labeled data from the source domain.However,in engineering scenarios,achieving such high-quality label annotation is difficult and expensive.The incorrect label annotation produces two negative effects:1)the complex decision boundary of diagnosis models lowers the generalization performance on the target domain,and2)the distribution of target domain samples becomes misaligned with the false-labeled samples.To overcome these negative effects,this article proposes a solution called the label recovery and trajectory designable network(LRTDN).LRTDN consists of three parts.First,a residual network with dual classifiers is to learn features from cross-domain samples.Second,an annotation check module is constructed to generate a label anomaly indicator that could modify the abnormal labels of false-labeled samples in the source domain.With the training of relabeled samples,the complexity of diagnosis model is reduced via semi-supervised learning.Third,the adaptation trajectories are designed for sample distributions across domains.This ensures that the target domain samples are only adapted with the pure-labeled samples.The LRTDN is verified by two case studies,in which the diagnosis knowledge of bearings is transferred across different working conditions as well as different yet related machines.The results show that LRTDN offers a high diagnosis accuracy even in the presence of incorrect annotation.展开更多
Recently,intelligent fault diagnosis based on deep learning has been extensively investigated,exhibiting state-of-the-art performance.However,the deep learning model is often not truly trusted by users due to the lack...Recently,intelligent fault diagnosis based on deep learning has been extensively investigated,exhibiting state-of-the-art performance.However,the deep learning model is often not truly trusted by users due to the lack of interpretability of“black box”,which limits its deployment in safety-critical applications.A trusted fault diagnosis system requires that the faults can be accurately diagnosed in most cases,and the human in the deci-sion-making loop can be found to deal with the abnormal situa-tion when the models fail.In this paper,we explore a simplified method for quantifying both aleatoric and epistemic uncertainty in deterministic networks,called SAEU.In SAEU,Multivariate Gaussian distribution is employed in the deep architecture to compensate for the shortcomings of complexity and applicability of Bayesian neural networks.Based on the SAEU,we propose a unified uncertainty-aware deep learning framework(UU-DLF)to realize the grand vision of trustworthy fault diagnosis.Moreover,our UU-DLF effectively embodies the idea of“humans in the loop”,which not only allows for manual intervention in abnor-mal situations of diagnostic models,but also makes correspond-ing improvements on existing models based on traceability analy-sis.Finally,two experiments conducted on the gearbox and aero-engine bevel gears are used to demonstrate the effectiveness of UU-DLF and explore the effective reasons behind.展开更多
Aerial scene recognition(ASR)has attracted great attention due to its increasingly essential applications.Most of the ASR methods adopt the multi‐scale architecture because both global and local features play great r...Aerial scene recognition(ASR)has attracted great attention due to its increasingly essential applications.Most of the ASR methods adopt the multi‐scale architecture because both global and local features play great roles in ASR.However,the existing multi‐scale methods neglect the effective interactions among different scales and various spatial locations when fusing global and local features,leading to a limited ability to deal with challenges of large‐scale variation and complex background in aerial scene images.In addition,existing methods may suffer from poor generalisations due to millions of to‐belearnt parameters and inconsistent predictions between global and local features.To tackle these problems,this study proposes a scale‐wise interaction fusion and knowledge distillation(SIF‐KD)network for learning robust and discriminative features with scaleinvariance and background‐independent information.The main highlights of this study include two aspects.On the one hand,a global‐local features collaborative learning scheme is devised for extracting scale‐invariance features so as to tackle the large‐scale variation problem in aerial scene images.Specifically,a plug‐and‐play multi‐scale context attention fusion module is proposed for collaboratively fusing the context information between global and local features.On the other hand,a scale‐wise knowledge distillation scheme is proposed to produce more consistent predictions by distilling the predictive distribution between different scales during training.Comprehensive experimental results show the proposed SIF‐KD network achieves the best overall accuracy with 99.68%,98.74%and 95.47%on the UCM,AID and NWPU‐RESISC45 datasets,respectively,compared with state of the arts.展开更多
基金the National Key R&D Program of China(2022YFB3402100)the National Science Fund for Distinguished Young Scholars of China(52025056)+4 种基金the National Natural Science Foundation of China(52305129)the China Postdoctoral Science Foundation(2023M732789)the China Postdoctoral Innovative Talents Support Program(BX20230290)the Open Foundation of Hunan Provincial Key Laboratory of Health Maintenance for Mechanical Equipment(2022JXKF JJ01)the Fundamental Research Funds for Central Universities。
文摘The success of deep transfer learning in fault diagnosis is attributed to the collection of high-quality labeled data from the source domain.However,in engineering scenarios,achieving such high-quality label annotation is difficult and expensive.The incorrect label annotation produces two negative effects:1)the complex decision boundary of diagnosis models lowers the generalization performance on the target domain,and2)the distribution of target domain samples becomes misaligned with the false-labeled samples.To overcome these negative effects,this article proposes a solution called the label recovery and trajectory designable network(LRTDN).LRTDN consists of three parts.First,a residual network with dual classifiers is to learn features from cross-domain samples.Second,an annotation check module is constructed to generate a label anomaly indicator that could modify the abnormal labels of false-labeled samples in the source domain.With the training of relabeled samples,the complexity of diagnosis model is reduced via semi-supervised learning.Third,the adaptation trajectories are designed for sample distributions across domains.This ensures that the target domain samples are only adapted with the pure-labeled samples.The LRTDN is verified by two case studies,in which the diagnosis knowledge of bearings is transferred across different working conditions as well as different yet related machines.The results show that LRTDN offers a high diagnosis accuracy even in the presence of incorrect annotation.
基金supported in part by the National Natural Science Foundation of China(52105116)Science Center for gas turbine project(P2022-DC-I-003-001)the Royal Society award(IEC\NSFC\223294)to Professor Asoke K.Nandi.
文摘Recently,intelligent fault diagnosis based on deep learning has been extensively investigated,exhibiting state-of-the-art performance.However,the deep learning model is often not truly trusted by users due to the lack of interpretability of“black box”,which limits its deployment in safety-critical applications.A trusted fault diagnosis system requires that the faults can be accurately diagnosed in most cases,and the human in the deci-sion-making loop can be found to deal with the abnormal situa-tion when the models fail.In this paper,we explore a simplified method for quantifying both aleatoric and epistemic uncertainty in deterministic networks,called SAEU.In SAEU,Multivariate Gaussian distribution is employed in the deep architecture to compensate for the shortcomings of complexity and applicability of Bayesian neural networks.Based on the SAEU,we propose a unified uncertainty-aware deep learning framework(UU-DLF)to realize the grand vision of trustworthy fault diagnosis.Moreover,our UU-DLF effectively embodies the idea of“humans in the loop”,which not only allows for manual intervention in abnor-mal situations of diagnostic models,but also makes correspond-ing improvements on existing models based on traceability analy-sis.Finally,two experiments conducted on the gearbox and aero-engine bevel gears are used to demonstrate the effectiveness of UU-DLF and explore the effective reasons behind.
基金supported in part by the National Natural Science Foundation of China under Grant 62201452,2271296 and 62201453in part by the Natural Science Basic Research Programme of Shaanxi under Grant 2022JQ‐592+1 种基金in part by the Special Construction Fund for Key Disciplines of Shaanxi Provincial Higher Education,in part by the Natural Science Basic Research Program of Shaanxi under Grant 2021JC‐47in part by Scientific Research Program Funded by Shaanxi Provincial Education Department under Grant 22JK0568.
文摘Aerial scene recognition(ASR)has attracted great attention due to its increasingly essential applications.Most of the ASR methods adopt the multi‐scale architecture because both global and local features play great roles in ASR.However,the existing multi‐scale methods neglect the effective interactions among different scales and various spatial locations when fusing global and local features,leading to a limited ability to deal with challenges of large‐scale variation and complex background in aerial scene images.In addition,existing methods may suffer from poor generalisations due to millions of to‐belearnt parameters and inconsistent predictions between global and local features.To tackle these problems,this study proposes a scale‐wise interaction fusion and knowledge distillation(SIF‐KD)network for learning robust and discriminative features with scaleinvariance and background‐independent information.The main highlights of this study include two aspects.On the one hand,a global‐local features collaborative learning scheme is devised for extracting scale‐invariance features so as to tackle the large‐scale variation problem in aerial scene images.Specifically,a plug‐and‐play multi‐scale context attention fusion module is proposed for collaboratively fusing the context information between global and local features.On the other hand,a scale‐wise knowledge distillation scheme is proposed to produce more consistent predictions by distilling the predictive distribution between different scales during training.Comprehensive experimental results show the proposed SIF‐KD network achieves the best overall accuracy with 99.68%,98.74%and 95.47%on the UCM,AID and NWPU‐RESISC45 datasets,respectively,compared with state of the arts.