Pith. sign in

REVIEW 3 major objections 5 minor 44 references

Integrating Frequency Guidance into Multi-source Domain Generalization for Bearing Fault Diagnosis

T0 review · 3 major / 5 minor · reviewed 2026-08-09 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read FARNet claims that separating Fourier phase from amplitude lets a bearing fault diagnosis model generalize to unseen loads and speeds, with 84.01% average accuracy on CWRU and 82.32% on SJTU.

desk verdict A plausible frequency-based augmentation architecture for bearing-fault DG whose reported gains are undermined by hyperparameters selected on the target test tasks. read the letter →

arxiv 2502.00545 v1 pith:PJCLN5XV submitted 2025-02-01 cs.LG cs.AIcs.CV

Xiaotong Tu , Chenyu Ma , Qingyao Wu , Yinhao Liu , Hongyang Zhang This is my paper · ORCID
classification cs.LGcs.AIcs.CV
keywords bearingfaultdiagnosisdomaingeneralizationFouriertransformdataaugmentationmanifoldtripletlossfrequency-spatialinteractionmulti-sourcephaseinvariance
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to establish that a fault-diagnosis model can generalize from known working conditions to unseen ones by using what the Fourier transform separates: phase as the stable, category-carrying part of a vibration signal and amplitude as the changing, domain-specific style. It proposes FARNet, a network with an amplitude sub-network and a phase sub-network that reconstruct augmented fault domains, a Frequency-Spatial Interaction Module that fuses global frequency information with local spatial features, and a manifold triplet loss that tightens class boundaries in a learned metric space. On the CWRU rolling-bearing benchmark the reported average accuracy is 84.01%, and on the SJTU rotor benchmark it is 82.32%, above the compared domain-adaptation and domain-generalization methods. If the claim holds, industrial monitoring systems would need labeled examples from fewer working conditions to deploy reliable fault diagnosis on previously unseen loads, speeds, and fault sizes.

What carries the argument

The load-bearing object is the Fourier-based Augmentation Reconstruction Network (FARNet), a pair of encoder-decoder sub-networks that operate on the amplitude spectrum and the phase spectrum of the input vibration signal. The amplitude sub-network is trained by Eq. (3) so that the amplitude of its output $A(X_{\mathrm{out1}})$ matches the ground-truth source amplitude $A(X_{\mathrm{gt}})$, and the phase sub-network is trained by Eq. (4) so that the phase of its output $P(X_{\mathrm{out2}})$ matches $P(X_{\mathrm{gt}})$; together these form the augmentation loss in Eq. (5). Inside both sub-networks, the Frequency-Spatial Interaction Module (FSIM) alternates a Fourier-domain branch, which processes amplitude or phase after a $1\times1$ convolution, with a spatial residual branch, and fuses the two through $3\times3$ convolutions. The manifold triplet loss in Eqs. (8)-(9) replaces the Euclidean distance with a piecewise-linear, nonlinear-activated distance $d_{\mathrm{new}}(x)$ that scales long distances up and short distances down, mining harder positive and negative samples in a manifold space.

What would settle it

Measure a distributional distance such as maximum mean discrepancy between Fourier phase distributions of the same fault class across different loads, speeds, and fault sizes on CWRU and SJTU; if phase distances are not systematically smaller than amplitude distances, the phase-invariance premise fails and FARNet's reconstruction objective loses its justification. A second check is to re-run the released code to confirm whether the SJTU average is 82.23% or 82.32%.

Watch

Extended reading notes

Core claim

The central claim is that the Fourier phase and amplitude of bearing vibration signals encode different kinds of information: phase carries category structure that stays aligned across domains, while amplitude carries stylistic, domain-specific variation. FARNet exploits this by reconstructing augmented fault domains, training an amplitude sub-network to match the amplitude of a selected ground-truth source domain and a phase sub-network to match its phase, so the model sees synthetic intermediate domains during training. A Frequency-Spatial Interaction Module (FSIM) lets each sub-network combine global frequency-domain information with local convolutional features, and a manifold triplet loss, built on a leaky-ReLU-like distance that breaks Euclidean triangle inequalities, pulls same-class features together and pushes different classes apart. The paper reports that this combination reaches 84.01% average accuracy on CWRU and 82.32% on SJTU in multi-source domain generalization settings, outperforming the compared domain-adaptation and domain-generalization baselines.

Load-bearing premise

The whole method rests on the assumption, supported mainly by the T-SNE plot in Fig. 1, that the Fourier phase of bearing vibration signals is domain-invariant while the amplitude is domain-specific; if that split is wrong, the reconstructed synthetic domains could inject the wrong information and the reported gains would not carry to new working conditions.

Editorial extensions

If this is right

  • If the reported results hold, a model trained on two or three known working conditions can diagnose faults under unseen loads, speeds, and fault sizes without any target-domain data.
  • Frequency-domain augmentation by phase and amplitude reconstruction can be added on top of an existing recognition backbone such as ResNet18, with only the augmentation losses and the manifold triplet loss as extra training objectives.
  • The ablation results indicate that the frequency augmentation module itself, rather than metric learning alone, drives most of the accuracy gain over the ResNet18 baseline.
  • The low standard deviation reported on the CWRU tasks (about 0.98 percentage points around the 84.01% average) suggests the method is stable across different choices of which source domains are given.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Beyond the paper's experiments, if Fourier phase is genuinely domain-invariant across loads, speeds, and sensor positions, the same reconstruction module could transfer to other rotating-machinery fault datasets without architectural changes; the paper does not test this.
  • The paper's evidence for phase invariance is a single T-SNE visualization, so a quantitative distributional distance comparison across many domains would be a natural next check of the load-bearing premise.
  • Editorial flag: the abstract reports 82.32% average accuracy on SJTU while Table 2 reports 82.23% for the same entry, and the paper does not reconcile the discrepancy.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The manuscript proposes FARNet, a Fourier-based augmentation reconstruction network for multi-source domain generalization in bearing fault diagnosis. The method decomposes the Fourier spectrum into amplitude and phase components, uses two sub-networks (with a Frequency-Spatial Interaction Module, FSIM) to reconstruct amplitude and phase against a selected source ground-truth domain, and adds a "manifold triplet loss" with a nonlinear distance to the classification objective. Experiments are reported on the CWRU and SJTU datasets under leave-one-domain-out tasks. The paper claims state-of-the-art average accuracies of 84.01% (CWRU) and 82.32% (SJTU, with 82.23% appearing in Table 2), and presents ablations of the augmentation module, the triplet loss, and the manifold triplet loss.

Significance. The paper addresses a practically relevant problem and combines several plausible components: Fourier-domain augmentation, frequency-spatial interaction, and hard-sample mining. The authors provide extensive ablations, run each experiment five times, and state that code will be released. If the evaluation protocol is corrected, the approach could be a useful benchmark contribution. However, as presented, the central domain-generalization claim is weakened by hyperparameter selection on the target test tasks and by qualitative-only support for the phase-invariance premise.

major comments (3)
  1. [§4.2, §4.4.2, §4.4.3, Figs. 8–9] The multi-source domain generalization claim is undermined by the hyperparameter selection protocol. The authors state that λ2/λ1 = 2 and k = 3 were selected because they give "optimal fault diagnosis accuracy" on the two benchmark test sets, and §4.2 describes no held-out validation split. Since the target domains are supposed to be unseen, selecting hyperparameters on the target test tasks makes the averages in Table 2 partly fitted to the benchmarks and violates the DG protocol. Please introduce a validation split drawn from the source domains (or pre-specify hyperparameters before seeing target data), describe the selection procedure, and report final accuracies under that procedure. Without this, the claim of "superior results compared to current cross-domain approaches" is not supported.
  2. [§3.1, Fig. 1] The premise that Fourier phase is domain-invariant while amplitude carries domain-specific style is supported only by a qualitative T-SNE visualization of one fault class (IRF) on CWRU. This premise motivates the reconstruction losses in Eqs. (3)–(5) and the design of the amplitude/phase sub-networks. Please provide quantitative evidence across all fault categories and both datasets—for example, distribution distances between amplitude and phase features across source–target pairs—and state whether the same separation holds under speed changes in SJTU. If the phase is not actually invariant, the interpretation of the augmentation module as transferring "phase semantics" is questionable.
  3. [§4.4.1, Table 3] The ablation text appears twice with inconsistent numbers: one version reports gains "up to 28.85% and 28.71%" on the two datasets, while the other reports "24.43% and 17.67%". The averages in Table 3 imply M2−M1 gains of 18.43 percentage points on CWRU and 24.91 on SJTU, so the reported boosts are not reproducible from the table. This must be corrected because the component-wise attribution of gains is a core part of the method's validation.
minor comments (5)
  1. [Abstract, §4.3, Table 2] The SJTU average accuracy is reported as 82.32% in the abstract and in §4.3, but Table 2 shows 82.23% for the average; please reconcile these numbers.
  2. [§3.4, Eq. (8)] The term "manifold triplet loss" is not justified by the definition: d_new(x) in Eq. (8) is a piecewise-linear rescaling of Euclidean distances, and the paper only states that it breaks the triangle inequality. Please clarify the manifold interpretation or rename the loss to avoid overclaiming the geometric mechanism.
  3. [§4.3] The text refers to "MDD" when the compared method is listed as "MMD" in Table 2; please correct the typo.
  4. [Fig. 3 caption] The caption ends with the incomplete phrase "while ."; the sentence should be completed so the described training flow is clear.
  5. [References] There are duplicated references for the same paper: entries [27], [32], and [34] all cite Ragab et al. Conditional Contrastive Domain Generalization; please consolidate them.

Circularity Check

1 steps flagged · score 6.0 of 10

Final generalization accuracies are selected, not predicted: λ2/λ1 and k are tuned on the exact CWRU/SJTU target tasks whose averages are then reported as FARNet's out-of-distribution performance.

  1. fitted input called prediction [Section 4.4.2 and 4.4.3 (hyperparameter ablation), with result reported in Section 4.3 / Table 2]
    "Through multiple experiments, the results, as depicted in Fig.8, reveal that our model achieves the optimal fault diagnosis accuracy on both datasets when the ratio of λ2/λ1 is set to 2. ... As depicted in Fig.9, the model achieves the best performance when k is set to 3. ... It can be observed from Table 2 that the final average accuracy of our FARNet method reaches the highest average accuracy of 84.01% on the CWRU dataset and 82.32% on the SJTU dataset, which are superior to the comparison methods."

    The reported generalization numbers are produced by selecting two hyperparameters (λ2/λ1=2 and k=3) to maximize 'fault diagnosis accuracy on both datasets'—the same two benchmark target tasks on which Table 2 then evaluates FARNet. No held-out validation split is described in Section 4.2; Section 4.2 simply refers to Sections 4.4.2 and 4.4.3 for the loss hyper-parameters. Therefore the 84.01% and 82.32% averages are not independent out-of-distribution predictions but the maximum over the searched configurations on the test tasks themselves. The central claim of superiority over cross-domain baselines is partly forced by this test-time model selection; the comparison methods were not given the same target-test hyperparameter search.

full rationale

The paper is an empirical architecture/loss proposal, so there is no analytical derivation chain whose equations collapse into their inputs. The reconstruction losses (Eqs. 3-5), FSIM, and manifold triplet loss are stated explicitly and do not reduce to each other. The phase-invariance premise is supported only by a qualitative T-SNE (Fig. 1) and is weak evidence, but that is an under-support problem, not circularity. The main circularity is in the evaluation protocol: Section 4.4.2 chooses λ2/λ1=2 because it gives 'the optimal fault diagnosis accuracy on both datasets' and Section 4.4.3 chooses k=3 because it gives 'the best performance' on the same datasets; these exact tasks are then used in Table 2 to claim state-of-the-art 'superior results'. Thus the headline accuracy is a selected maximum over the searched hyperparameters on the target test sets rather than an unbiased generalization estimate. This does not invalidate the architecture, but it makes the central empirical claim partially self-fulfilling. No load-bearing self-citation or uniqueness-importation issue is present.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The central claim rests on the domain assumption of phase/amplitude semantic separation, the design assumption that matching spectra to one source domain creates useful synthetic domains, and several hand-tuned hyperparameters. No new physical entities are introduced.

free parameters (6)
  • k = 3
    Scale factor in manifold triplet distance Eq. (8); selected by test-set ablation (Fig. 9) which tunes on the target tasks.
  • lambda1 = 0.1
    Weight for amplitude reconstruction loss; lambda2/lambda1 ratio chosen by test-set ablation (Fig. 8).
  • lambda2 = 0.2
    Weight for phase reconstruction loss; selected jointly with lambda1 in Fig. 8.
  • alpha = 0.01
    Weight for manifold triplet loss in Eq. (12); chosen by hand without ablation.
  • gamma = 0.3
    Triplet margin in Eq. (9); chosen by hand without ablation.
  • r = mean batch distance
    Threshold in Eq. (8); computed per batch from data rather than tuned.
assumptions (4)
  • domain assumption Fourier phase preserves high-level semantic and category information and is relatively domain-invariant, while amplitude carries style and domain-specific statistics.
    Motivates FARNet; supported qualitatively by T-SNE in Fig. 1 and cited vision literature, not quantitatively validated for bearing signals.
  • domain assumption Reconstructing amplitude and phase of source signals to match a chosen reference source domain creates useful synthetic domains for domain generalization.
    Defines Lamp and Lpha in Eqs. (3)-(5); there is no proof that this augmentation reduces true domain shift.
  • ad hoc to paper The leaky-relu-like distance in Eq. (8), which breaks the triangle inequality, yields a meaningful manifold metric that improves hard-sample mining.
    No formal definition or proof is given; the manifold terminology is used without a manifold construction.
  • standard math Fourier transform and inverse Fourier transform properties are standard.
    Used throughout Section 3 to separate and recombine amplitude and phase.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Integrating Frequency Guidance into Multi-source Domain Generalization for Bearing Fault Diagnosis." pith.science (2026). https://pith.science/paper/PJCLN5XV

@misc{pith2026250200545,
  author       = {Pith},
  title        = {Pith review of: Integrating Frequency Guidance into Multi-source Domain Generalization for Bearing Fault Diagnosis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PJCLN5XV}},
  note         = {Machine review of arXiv:2502.00545}
}
read the original abstract

Recent generalizable fault diagnosis researches have effectively tackled the distributional shift between unseen working conditions. Most of them mainly focus on learning domain-invariant representation through feature-level methods. However, the increasing numbers of unseen domains may lead to domain-invariant features contain instance-level spurious correlations, which impact the previous models' generalizable ability. To address the limitations, we propose the Fourier-based Augmentation Reconstruction Network, namely FARNet.The methods are motivated by the observation that the Fourier phase component and amplitude component preserve different semantic information of the signals, which can be employed in domain augmentation techniques. The network comprises an amplitude spectrum sub-network and a phase spectrum sub-network, sequentially reducing the discrepancy between the source and target domains. To construct a more robust generalized model, we employ a multi-source domain data augmentation strategy in the frequency domain. Specifically, a Frequency-Spatial Interaction Module (FSIM) is introduced to handle global information and local spatial features, promoting representation learning between the two sub-networks. To refine the decision boundary of our model output compared to conventional triplet loss, we propose a manifold triplet loss to contribute to generalization. Through extensive experiments on the CWRU and SJTU datasets, FARNet demonstrates effective performance and achieves superior results compared to current cross-domain approaches on the benchmarks.

Figures

Figures reproduced from arXiv: 2502.00545 by the authors.

Figure 1
Figure 1. The T-SNE[3] of amplitude and phase of category inner race fault (IRF) with 7 mils, 14 mils, and 21 mils (three domains) on CWRU dataset are depicted on (a) and (b), respectively. It’s obvious that amplitude features are compose of domain-relevant information and phase features are compose of domain-invariant information, which can help to obtain more diverse data for domain generalization. models lack access to unl… view at source ↗
Figure 2
Figure 2. The data distribution of a synthetic multi-class dataset in Euclidean space and Manifold space in [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. The overall framework proposed in this paper consists of two main modules: the The Fourier-based Augmentation Reconstruction [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: The illustration depicts the amplitude format of the Frequency-Spatial Interaction Module (FSIM). The module comprises two flows: [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: (a) CWRU experimental setup for machinery fault diagnosis [ [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: Picture of rotor crack in SJTU dataset [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: Confusion matrices of different diagnostic methods under SJ1 2→SJ3 condition of SJTU dataset.The first line is the classical domain adaptation, domain generalization approach and the second line is a domain generalization approach for bearing fault diagnosis [PITH_FUL…
Figure 8
Figure 8. Figure 8: The ablation experiments of hyperparameter [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]
Figure 9
Figure 9. Figure 9: The ablation experiments of hyperparameter k for [PITH_FULL_IMAGE:figures/full_fig_p013_9.png]
Figure 10
Figure 10. Figure 10: The T-SNE results of first and second categories from the four domains (1000 and 1200 are real domains, while 1000 [PITH_FULL_IMAGE:figures/full_fig_p014_10.png]
Figure 11
Figure 11. Figure 11: T-SNE of M2, M3, and M4 diagnostic methods under SJ [PITH_FULL_IMAGE:figures/full_fig_p014_11.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 42 canonical work pages

  1. [42]

    Ragab, Z

    M. Ragab, Z. Chen, W. Zhang, E. Eldele, M. Wu, C.-K. Kwoh, X. Li, Conditional contrastive domain generalization for fault diagnosis, IEEE Transactions on Instrumentation and Measurement 71 (2022) 1–12

    work page 2022

  2. [1]

    Y . Liu, Y . Wang, T. W. Chow, B. Li, Deep adversarial subdomain adaptation network for intelligent fault diagnosis, IEEE Transactions on Industrial Informatics 18 (9) (2022) 6038–6046

    work page 2022

  3. [2]

    Z. Yang, Z. Ge, On paradigm of industrial big data analytics: From evolution to revolution, IEEE Transactions on Industrial Informatics 18 (12) (2022) 8373–8388

    work page 2022

  4. [3]

    Van der Maaten, G

    L. Van der Maaten, G. Hinton, Visualizing data using t-sne., Journal of machine learning research 9 (11) (2008)

    2008

  5. [4]

    L. Cui, Z. Dong, H. Xu, D. Zhao, Triplet attention-enhanced residual tree-inspired decision network: A hierarchical fault diagnosis model for unbalanced bearing datasets, Advanced Engineering Informatics 59 (2024) 102322

    work page 2024

  6. [5]

    D. Ruan, J. Wang, J. Yan, C. G ¨uhmann, Cnn parameter design based on fault signal analysis and its application in bearing fault diagnosis, Advanced Engineering Informatics 55 (2023) 101877

    work page 2023

  7. [6]

    M. He, D. He, Deep learning based approach for bearing fault diagnosis, IEEE Transactions on Industry Applications 53 (3) (2017) 3057– 3065

    work page 2017

  8. [7]

    X. Chen, R. Yang, Y . Xue, M. Huang, R. Ferrero, Z. Wang, Deep transfer learning for bearing fault diagnosis: A systematic review since 2016, IEEE Transactions on Instrumentation and Measurement 72 (2023) 1–21. 14

    work page 2023

Show all 44 references
  1. [8]

    Y . Xiao, H. Shao, S. Han, Z. Huo, J. Wan, Novel joint transfer network for unsupervised bearing fault diagnosis from simulation domain to experimental domain, IEEE/ASME Transactions on Mechatronics 27 (6) (2022) 5254–5263

    2022

  2. [9]

    Y . Li, Y . Yang, W. Zhou, T. Hospedales, Feature-critic networks for heterogeneous domain generalization, in: International Conference on Machine Learning, PMLR, 2019, pp. 3915–3924

    2019

  3. [10]

    L. Ren, T. Mo, X. Cheng, Meta-learning based domain generalization framework for fault diagnosis with gradient aligning and semantic matching, IEEE Transactions on Industrial Informatics 20 (1) (2023) 754–764

    2023

  4. [11]

    Krueger, E

    D. Krueger, E. Caballero, J.-H. Jacobsen, A. Zhang, J. Binas, D. Zhang, R. Le Priol, A. Courville, Out-of-distribution generalization via risk extrapolation (rex), in: International Conference on Machine Learning, PMLR, 2021, pp. 5815–5826

    2021

  5. [12]

    J. Chen, R. Huang, K. Zhao, W. Wang, L. Liu, W. Li, Multiscale convolutional neural network with feature alignment for bearing fault diagnosis, IEEE Transactions on Instrumentation and Measurement 70 (2021) 1–10

    2021

  6. [14]

    Y . Shi, A. Deng, M. Deng, M. Xu, Y . Liu, X. Ding, W. Bian, Domain augmentation generalization network for real-time fault diagnosis under unseen working conditions, Reliability Engineering & System Safety 235 (2023) 109188

    2023

  7. [15]

    K. Zhou, Y . Yang, T. Hospedales, T. Xiang, Learning to generate novel domains for domain generalization, in: Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XVI 16, Springer, 2020, pp. 561–578

    2020

  8. [16]

    V olpi, H

    R. V olpi, H. Namkoong, O. Sener, J. C. Duchi, V . Murino, S. Savarese, Generalizing to unseen domains via adversarial data augmentation, Advances in neural information processing systems 31 (2018)

    2018

  9. [17]

    K. Zhou, Y . Yang, T. Hospedales, T. Xiang, Deep domain-adversarial image generation for domain generalisation, in: Proceedings of the AAAI conference on artificial intelligence, V ol. 34, 2020, pp. 13025–13032

    2020

  10. [18]

    Huang, Y

    J. Huang, Y . Liu, F. Zhao, K. Yan, J. Zhang, Y . Huang, M. Zhou, Z. Xiong, Deep fourier-based exposure correction network with spatial- frequency interaction, in: European Conference on Computer Vision, Springer, 2022, pp. 163–180

    2022

  11. [19]

    A. V . Oppenheim, J. S. Lim, The importance of phase in signals, Proceedings of the IEEE 69 (5) (1981) 529–541

    1981

  12. [20]

    W. A. Smith, R. B. Randall, Rolling element bearing diagnostics using the case western reserve university data: A benchmark study, Mechan- ical systems and signal processing 64 (2015) 100–131

    2015

  13. [21]

    C. Liu, X. Wang, S. Li, Y . Wang, X. Qian, Fsi: Frequency and spatial interactive learning for image restoration in under-display cameras, in: Proceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 12537–12546

    2023

  14. [22]

    C. Wang, J. Jiang, Z. Zhong, X. Liu, Spatial-frequency mutual learning for face super-resolution, in: Proceedings of the IEEE /CVF Confer- ence on Computer Vision and Pattern Recognition, 2023, pp. 22356–22366

    2023

  15. [23]

    F. Qu, J. Liu, X. Liu, L. Jiang, A multi-fault detection method with improved triplet loss based on hard sample mining, IEEE Transactions on Sustainable Energy 12 (1) (2020) 127–137

    2020

  16. [24]

    C. Wang, C. Xin, Z. Xu, A novel deep metric learning model for imbalanced fault diagnosis and toward open-set classification, Knowledge- Based Systems 220 (2021) 106925

    2021

  17. [25]

    C. Zhao, W. Shen, Imbalanced domain generalization via semantic-discriminative augmentation for intelligent fault diagnosis, Advanced Engineering Informatics 59 (2024) 102262

    2024

  18. [26]

    X. Wang, F. Liu, Triplet loss guided adversarial domain adaptation for bearing fault diagnosis, Sensors 20 (1) (2020) 320

    2020

  19. [28]

    Iscen, G

    A. Iscen, G. Tolias, Y . Avrithis, O. Chum, Mining on manifolds: Metric learning without labels, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018, pp. 7642–7651

    2018

  20. [29]

    Huang, Z

    K. Huang, Z. Ren, L. Zhu, T. Lin, Y . Zhu, L. Zeng, J. Wan, Intra-domain self generalization network for intelligent fault diagnosis of bearings under unseen working conditions, Advanced Engineering Informatics 64 (2025) 102997

    2025

  21. [30]

    Z. Mo, Z. Zhang, K.-L. Tsui, Distance aware risk minimization for domain generalization in machine fault diagnosis, IEEE Internet of Things Journal (2024)

    2024

  22. [31]

    C. Zhao, W. Shen, A domain generalization network combing invariance and specificity towards real-time intelligent fault diagnosis, Me- chanical Systems and Signal Processing 173 (2022) 108990

    2022

  23. [33]

    Q. Li, L. Chen, L. Kong, D. Wang, M. Xia, C. Shen, Cross-domain augmentation diagnosis: An adversarial domain-augmented generalization method for fault diagnosis under unseen working conditions, Reliability Engineering & System Safety 234 (2023) 109171

    2023

  24. [35]

    Y . Yang, S. Soatto, Fda: Fourier domain adaptation for semantic segmentation, in: Proceedings of the IEEE /CVF conference on computer vision and pattern recognition, 2020, pp. 4085–4095

    2020

  25. [36]

    Q. Xu, R. Zhang, Y . Zhang, Y . Wang, Q. Tian, A fourier-based framework for domain generalization, in: Proceedings of the IEEE /CVF Conference on Computer Vision and Pattern Recognition, 2021, pp. 14383–14392

    2021

  26. [37]

    S. Lin, Z. Zhang, Z. Huang, Y . Lu, C. Lan, P. Chu, Q. You, J. Wang, Z. Liu, A. Parulkar, et al., Deep frequency filtering for domain generalization, in: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2023, pp. 11797–11807

    2023

  27. [38]

    Z. Fan, Q. Xu, C. Jiang, S. X. Ding, Deep mixed domain generalization network for intelligent fault diagnosis under unseen conditions, IEEE Transactions on Industrial Electronics 71 (1) (2023) 965–974

    2023

  28. [39]

    Katznelson, An introduction to harmonic analysis, Cambridge University Press, 2004

    Y . Katznelson, An introduction to harmonic analysis, Cambridge University Press, 2004

    2004

  29. [40]

    Hadsell, S

    R. Hadsell, S. Chopra, Y . LeCun, Dimensionality reduction by learning an invariant mapping, in: 2006 IEEE computer society conference on computer vision and pattern recognition (CVPR’06), V ol. 2, IEEE, 2006, pp. 1735–1742

    2006

  30. [41]

    W. A. Smith, R. B. Randall, Rolling element bearing diagnostics using the case western reserve university data: A benchmark study, Mechan- ical systems and signal processing 64 (2015) 100–131. 15

    2015

  31. [43]

    K. He, X. Zhang, S. Ren, J. Sun, Deep residual learning for image recognition, in: Proceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 770–778

    2016

  32. [44]

    Q. Qian, Y . Wang, T. Zhang, Y . Qin, Maximum mean square discrepancy: a new discrepancy representation metric for mechanical fault transfer diagnosis, Knowledge-Based Systems 276 (2023) 110748

    2023

  33. [45]

    X. Jin, C. Lan, W. Zeng, Z. Chen, L. Zhang, Style normalization and restitution for generalizable person re-identification, in: proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2020, pp. 3143–3152

    2020

  34. [46]

    Hongyi, C

    Z. Hongyi, C. Moustapha, D. Yann N., L.-P. David, mixup: Beyond empirical risk minimization, International Conference on Learning Representations (2018)

    2018

  35. [47]

    K. Zhou, Y . Yang, Y . Qiao, T. Xiang, Domain generalization with mixstyle, in: ICLR, 2021

    2021

  36. [48]

    Han, Y .-F

    T. Han, Y .-F. Li, M. Qian, A hybrid generalization network for intelligent fault diagnosis of rotating machinery under unseen working conditions, IEEE Transactions on Instrumentation and Measurement 70 (2021) 1–11. 16

    2021

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.