Yifei Ding
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Yifei Ding has a total of 31 co-authors. Here are the top co-authors by number of shared publications:
- Minping Jia
School of Mechanical Engineering, Southeast University, Nanjing, China
12 shared publications - Xiaoli Zhao
Nanjing University of Science and Technology
9 shared publications - Yudong Cao
Nanjing University of Aeronautics and Astronautics
8 shared publications - Chi-Guhn Lee
University of Toronto
7 shared publications - Peng Ding
Yangzhou University
7 shared publications Xiaoli Zhao
School of Mechanical Engineering, Nanjing University of Science and Technology, Nanjing 210094, People’s Republic of China
5 shared publications
Latest Publications
- 23 September 2026
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 37 (38) , 386105
Causality-inspired graph representation with adaptive thresholding for robust fault detection
Reliable fault detection in complex industrial systems requires capturing structured dependencies among process variables rather than relying solely on individual signal patterns. However, most existing approaches primarily depend on reconstruction or prediction errors, which may overlook relational interactions and lead to unstable detection performance under varying operating conditions. To address this limitation, this paper proposes a framework named causality-inspired graph representation with adaptive thresholding (CGRA). CGRA combines temporal causal representation to capture structured temporal dependencies from multivariate time series and graph attention representation to encode these relations into discriminative features for anomaly scoring. An adaptive thresholding strategy is further employed to obtain statistically reliable decision boundaries. Experimental results on multiple industrial datasets demonstrate consistent performance improvements over representative existing methods.
- 23 September 2026
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 37 (38) , 386105
Causality-inspired graph representation with adaptive thresholding for robust fault detection
Reliable fault detection in complex industrial systems requires capturing structured dependencies among process variables rather than relying solely on individual signal patterns. However, most existing approaches primarily depend on reconstruction or prediction errors, which may overlook relational interactions and lead to unstable detection performance under varying operating conditions. To address this limitation, this paper proposes a framework named causality-inspired graph representation with adaptive thresholding (CGRA). CGRA combines temporal causal representation to capture structured temporal dependencies from multivariate time series and graph attention representation to encode these relations into discriminative features for anomaly scoring. An adaptive thresholding strategy is further employed to obtain statistically reliable decision boundaries. Experimental results on multiple industrial datasets demonstrate consistent performance improvements over representative existing methods.
- 24 July 2026
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 37 (30) , 305101
LLM-informed stage-aware and physics-guided remaining useful life prediction for rolling bearings
Remaining useful life (RUL) prediction of rolling bearings is a fundamental task in prognostics and health management. Existing data-driven approaches mainly rely on static training objectives and purely signal-level learning, which limits their ability to adapt to different degradation stages and to incorporate high-level degradation knowledge. An large language model (LLM)-informed RUL prediction framework is proposed, integrating LLM-based degradation reasoning, stage-aware adaptive learning, and physics-guided constraint regularization. Instead of directly processing raw vibration signals, the LLM is employed as a degradation reasoning module that infers structured degradation priors, including degradation stage, degradation trend, and physically meaningful constraints, from semantic health representations. These priors are subsequently embedded into a task-specific neural predictor through a stage-aware adaptive loss updating strategy and physics-inspired constraint losses. By explicitly modeling stage-dependent learning objectives and enforcing physically consistent degradation behavior, the proposed framework achieves more reliable lifecycle-wide RUL prediction. Experimental results demonstrate that the proposed method consistently outperforms existing data-driven approaches in terms of prediction accuracy and stability.
- 24 July 2026
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 37 (30) , 305101
LLM-informed stage-aware and physics-guided remaining useful life prediction for rolling bearings
Remaining useful life (RUL) prediction of rolling bearings is a fundamental task in prognostics and health management. Existing data-driven approaches mainly rely on static training objectives and purely signal-level learning, which limits their ability to adapt to different degradation stages and to incorporate high-level degradation knowledge. An large language model (LLM)-informed RUL prediction framework is proposed, integrating LLM-based degradation reasoning, stage-aware adaptive learning, and physics-guided constraint regularization. Instead of directly processing raw vibration signals, the LLM is employed as a degradation reasoning module that infers structured degradation priors, including degradation stage, degradation trend, and physically meaningful constraints, from semantic health representations. These priors are subsequently embedded into a task-specific neural predictor through a stage-aware adaptive loss updating strategy and physics-inspired constraint losses. By explicitly modeling stage-dependent learning objectives and enforcing physically consistent degradation behavior, the proposed framework achieves more reliable lifecycle-wide RUL prediction. Experimental results demonstrate that the proposed method consistently outperforms existing data-driven approaches in terms of prediction accuracy and stability.
- Open Access
- 8 April 2026
- Published by Oxford University Press (OUP) in Transportation Safety and Environment
- Vol. 8 (2)
Generalizable fault detection and diagnosis of bearings via time-frequency embedded pre-training
Deep learning has demonstrated significant potential in data-driven fault detection and diagnosis (FDD). However, its effectiveness often depends on large amounts of labeled data, which are costly and impractical to obtain in real-world industrial applications such as elevator systems. To address this limitation, this paper proposes a novel time-frequency embedded pre-training network (TFEPTN) that employs self-supervised learning to extract meaningful representations from unlabeled vibration signals. TFEPTN features a dual-branch Transformer-based encoder for time-domain and frequency-domain feature extraction, guided by a time-frequency consistency constraint to enhance the learning of domain-representative features. The model is first pre-trained in an unsupervised manner and then fine-tuned for downstream FDD tasks. Experimental results on datasets of rolling bearings demonstrate that TFEPTN achieves strong generalization and outperforms several state-of-the-art methods under limited supervision and cross-domain conditions.
- 8 April 2026
- Published by Oxford University Press (OUP) in Transportation Safety and Environment
- Vol. 8 (2)
Open AccessGeneralizable fault detection and diagnosis of bearings via time-frequency embedded pre-training
Deep learning has demonstrated significant potential in data-driven fault detection and diagnosis (FDD). However, its effectiveness often depends on large amounts of labeled data, which are costly and impractical to obtain in real-world industrial applications such as elevator systems. To address this limitation, this paper proposes a novel time-frequency embedded pre-training network (TFEPTN) that employs self-supervised learning to extract meaningful representations from unlabeled vibration signals. TFEPTN features a dual-branch Transformer-based encoder for time-domain and frequency-domain feature extraction, guided by a time-frequency consistency constraint to enhance the learning of domain-representative features. The model is first pre-trained in an unsupervised manner and then fine-tuned for downstream FDD tasks. Experimental results on datasets of rolling bearings demonstrate that TFEPTN achieves strong generalization and outperforms several state-of-the-art methods under limited supervision and cross-domain conditions.
- 19 March 2026
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 37 (11) , 116001
Neural stochastic differential equation based lightweight diffusion model for imbalanced fault diagnosis in rotating machinery
The issue of imbalanced data in fault diagnosis significantly degrades diagnostic performance for rotating machinery. Existing data augmentation methods have inherent limitations, with generative adversarial network exhibiting unstable training and susceptibility to pattern collapse. The generation quality of denoising diffusion probabilistic model is constrained by the accuracy of discretization modeling, and their training efficiency remains low. A diffuse autoencoder with the neural stochastic differential equation (SDE) diffusion process for data augmentation task is introduced in this paper. Firstly, a novel network architecture of encoder-diffusion-decoder method is proposed, which compresses the original signal into a compact feature representation in a low dimensional latent space through an encoder, laying a low complexity modeling foundation for subsequent diffusion processes. Secondly, lightweight neural networks are first used to parameterize the drift and diffusion terms in the SDE reverse process, enhancing the model’s ability to fit complex data distributions. Finally, a position-encoded temporal embedding strategy is introduced to reverse diffusion progress, which can ensure temporal coordination among the drift network, diffusion network, and Score network, providing a key guarantee for accurate modeling of the diffusion process. Experimental results demonstrate that proposed model can efficiently train and generate high-quality fault signals, and efficiently enhance the fault diagnosis performance of rotating equipment under imbalanced data.
- 19 March 2026
- Published by IOP Publishing in Measurement Science and Technology
- Vol. 37 (11) , 116001
Neural stochastic differential equation based lightweight diffusion model for imbalanced fault diagnosis in rotating machinery
The issue of imbalanced data in fault diagnosis significantly degrades diagnostic performance for rotating machinery. Existing data augmentation methods have inherent limitations, with generative adversarial network exhibiting unstable training and susceptibility to pattern collapse. The generation quality of denoising diffusion probabilistic model is constrained by the accuracy of discretization modeling, and their training efficiency remains low. A diffuse autoencoder with the neural stochastic differential equation (SDE) diffusion process for data augmentation task is introduced in this paper. Firstly, a novel network architecture of encoder-diffusion-decoder method is proposed, which compresses the original signal into a compact feature representation in a low dimensional latent space through an encoder, laying a low complexity modeling foundation for subsequent diffusion processes. Secondly, lightweight neural networks are first used to parameterize the drift and diffusion terms in the SDE reverse process, enhancing the model’s ability to fit complex data distributions. Finally, a position-encoded temporal embedding strategy is introduced to reverse diffusion progress, which can ensure temporal coordination among the drift network, diffusion network, and Score network, providing a key guarantee for accurate modeling of the diffusion process. Experimental results demonstrate that proposed model can efficiently train and generate high-quality fault signals, and efficiently enhance the fault diagnosis performance of rotating equipment under imbalanced data.
- 10 January 2026
- Published by Elsevier in Measurement
- Vol. 265
UFOCC: Contamination-Robust few-shot one-class learning for online detection of Bearing incipient faults
- 10 January 2026
- Published by Elsevier in Measurement
- Vol. 265
UFOCC: Contamination-Robust few-shot one-class learning for online detection of Bearing incipient faults
- 25 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74 (00189456) , 1-10
Robust Weakly Supervised Bearing Fault Diagnosis via Dual Learning and Curriculum Strategies
Deep transfer learning (DTL), especially domain adaptation (DA) based techniques, has broadly extended the scope of practical fault diagnosis which struggles with generalization under variable operating conditions. However, real-world applications often involve noisy source domains with significant label and feature corruption, posing severe challenges to existing DTL approaches. In this work, we introduce the weakly-supervised domain adaptation network (WSDAN), a novel framework designed to tackle these issues through transfer curriculum learning (TCL) and dual learning strategies. WSDAN prioritizes informative samples from noisy source domains and enhances cross-domain generalization by utilizing inherent supervision in unlabeled target domain data. Our extensive experiments demonstrate that WSDAN outperforms current DTL and DA methods, particularly in high noise and significant domain discrepancy scenarios. This study highlights WSDAN’s robust performance and practical utility, paving the way for more reliable fault diagnosis in complex, real-world conditions.
- 25 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74 (00189456) , 1-10
Robust Weakly Supervised Bearing Fault Diagnosis via Dual Learning and Curriculum Strategies
Deep transfer learning (DTL), especially domain adaptation (DA) based techniques, has broadly extended the scope of practical fault diagnosis which struggles with generalization under variable operating conditions. However, real-world applications often involve noisy source domains with significant label and feature corruption, posing severe challenges to existing DTL approaches. In this work, we introduce the weakly-supervised domain adaptation network (WSDAN), a novel framework designed to tackle these issues through transfer curriculum learning (TCL) and dual learning strategies. WSDAN prioritizes informative samples from noisy source domains and enhances cross-domain generalization by utilizing inherent supervision in unlabeled target domain data. Our extensive experiments demonstrate that WSDAN outperforms current DTL and DA methods, particularly in high noise and significant domain discrepancy scenarios. This study highlights WSDAN’s robust performance and practical utility, paving the way for more reliable fault diagnosis in complex, real-world conditions.
- 24 May 2024
- Published by Elsevier in Knowledge-Based Systems
- Vol. 299
Deep temporal–spectral domain adaptation for bearing fault diagnosis
- 24 May 2024
- Published by Elsevier in Knowledge-Based Systems
- Vol. 299
Deep temporal–spectral domain adaptation for bearing fault diagnosis
- 12 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Informatics
- Vol. 20 (2) , 1313-1323
Unsupervised Fault Detection With Deep One-Class Classification and Manifold Distribution Alignment
Fault detection or anomaly detection relies heavily on learning from datasets where only normal samples are available, resulting in the emergence of numerous one-class classification (OCC) methods. However, learning discriminative deep representatives with good generalization from cross-domain positive samples remains challenging. Therefore, this work proposes an end-to-end framework, deep transfer one-class classification (DTOCC) for unsupervised fault detection, which combines adversarial generative OCC and distribution alignment from the perspective of manifold learning. Specifically, pseudo-negative samples are generated outside the positive manifold, facilitating the model to learn discrimination with respect to normal and anomaly. Further, cross-domain positive samples are aligned in log-Euclidean manifold space to enhance representation learning. Then, we provide the specific implementations for fault detection and validate its superiority through case studies on multiclass and run-to-failure datasets, simulating both offline and online scenarios.
- 12 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Informatics
- Vol. 20 (2) , 1313-1323
Unsupervised Fault Detection With Deep One-Class Classification and Manifold Distribution Alignment
Fault detection or anomaly detection relies heavily on learning from datasets where only normal samples are available, resulting in the emergence of numerous one-class classification (OCC) methods. However, learning discriminative deep representatives with good generalization from cross-domain positive samples remains challenging. Therefore, this work proposes an end-to-end framework, deep transfer one-class classification (DTOCC) for unsupervised fault detection, which combines adversarial generative OCC and distribution alignment from the perspective of manifold learning. Specifically, pseudo-negative samples are generated outside the positive manifold, facilitating the model to learn discrimination with respect to normal and anomaly. Further, cross-domain positive samples are aligned in log-Euclidean manifold space to enhance representation learning. Then, we provide the specific implementations for fault detection and validate its superiority through case studies on multiclass and run-to-failure datasets, simulating both offline and online scenarios.
- 15 December 2022
- Published by Elsevier in Knowledge-Based Systems
- Vol. 261
Domain generalization via adversarial out-domain augmentation for remaining useful life prediction of bearings under unseen conditions
- 15 December 2022
- Published by Elsevier in Knowledge-Based Systems
- Vol. 261
Domain generalization via adversarial out-domain augmentation for remaining useful life prediction of bearings under unseen conditions
- Top Cited Papers
- 12 October 2022
- Published by Elsevier in Reliability Engineering & System Safety
- Vol. 230
Deep imbalanced domain adaptation for transfer learning fault diagnosis of bearings under multiple working conditions
- 12 October 2022
- Published by Elsevier in Reliability Engineering & System Safety
- Vol. 230
Top Cited PapersDeep imbalanced domain adaptation for transfer learning fault diagnosis of bearings under multiple working conditions