Pith. sign in

REVIEW 4 major objections 4 minor 52 references

Semi-Supervised Federated Learning via Dual Contrastive Learning and Soft Labeling for Intelligent Fault Diagnosis

T0 review · 4 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read A semi-supervised federated framework that shares only compact prototypes can lift fault-diagnosis accuracy by 1.15% to 7.85% when just 10% of data are labeled.

desk verdict A genuinely new combination of TLAW weighting, dual contrastive losses, and momentum prototype aggregation, but the headline accuracy claim is not reproducible from the paper's own tables and the proof of Theorem 1 has an invalid step; the evaluation protocol also likely favors the proposed method via per-client fine-tuning. read the letter →

arxiv 2507.14181 v1 pith:DKJ63YG7 submitted 2025-07-12 cs.LG cs.AI

Yajiao Dai , Jun Li , Zhen Mei , Yiyang Ni , Shi Jin , Zengxiang Li , Sheng Guo , Wei Xiang This is my paper · ORCID
classification cs.LGcs.AI
keywords semi-supervisedfederatedlearningintelligentfaultdiagnosiscontrastiveprototypeaggregationpseudo-labelweightingLaplacedistributionlabelscarcitynon-IIDdata
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

The paper aims to establish that federated intelligent fault diagnosis can be made accurate when each industrial client holds very few labeled samples and large amounts of unlabeled, non-identically distributed sensor data. Its framework, SSFL-DCSL, combines pseudo-label semi-supervised learning with contrastive representation learning, and communicates only compact per-class prototypes instead of full models. A truncated-Laplace sample-weighting function suppresses unreliable pseudo-labels, while a dual contrastive loss aligns features locally and against global prototypes. On three public benchmark datasets and a factory-collected pump dataset, the method reports accuracy gains of 1.15% to 7.85% over state-of-the-art baselines at a 10% label rate, with a more than 99% reduction in per-round communication. If these results replicate, factories could build shared diagnosis models from mostly unlabeled data with far less expert labeling and bandwidth.

What carries the argument

The central object is the per-class prototype, the average embedding of all samples (labeled and pseudo-labeled) of a class within a client. The server fuses local prototypes as $\tilde{P}_j = \frac{1}{|\mathcal{N}_j|}\sum_{k \in \mathcal{N}_j}\frac{|D^k_j|}{N_j} P^k_j$, then applies momentum $P^{t+1} = \kappa P^{t} + (1-\kappa)\tilde{P}^{t+1}$. Clients pull back the global prototypes and use them in a global contrastive loss $L_{GC} = -\sum_{i=1}^C \log \frac{\exp(\mathrm{sim}(P_i,\tilde{P}_i)/\tau_i)}{\sum_{j \neq i}\exp(\mathrm{sim}(P_i,\tilde{P}_j)/\tau_j)}$, while the local contrastive loss selects positives and negatives by pseudo-labels and uses a dynamic temperature. The TLAW weight $\lambda(p)=\lambda_{\max}2b_t\phi(\max(p);\mu_t,b_t)$ if $\max(p)<\mu_t$, and $\lambda_{\max}$ otherwise, where $\phi(x;\mu,b)=\frac{1}{2b}e^{-|x-\mu|/b}$ is the Laplace density and $\mu_t,b_t$ are estimated by exponential moving average of batch confidence, controls how much each pseudo-label contributes to the unsupervised loss.

What would settle it

Run the strongest baselines with the same per-client fine-tuning and local test protocol described in Section III-A at a 10% label rate; if their accuracy matches SSFL-DCSL, the claim that the proposed components drive the gain is falsified. As a second test, replace the Laplace weighting with a fixed confidence threshold and check whether accuracy on the first benchmark drops by the reported 0.51–3.87 points.

Watch

Extended reading notes

Core claim

The central claim is that SSFL-DCSL, a federated semi-supervised framework, achieves state-of-the-art intelligent fault diagnosis under severe label scarcity and non-IID client data by combining three design choices: a truncated Laplace-based adaptive sample weighting (TLAW) function that down-weights low-confidence pseudo-labels, a dual contrastive loss made of a local sample-level contrast term and a global prototype-level alignment term, and prototype aggregation with momentum that replaces full-model communication with per-class feature averages. The authors report that with only 10% of data labeled, SSFL-DCSL improves accuracy by 1.15% to 7.85% over the compared state-of-the-art methods across three public benchmark datasets and a chemical-plant pump dataset, while reducing per-round communication by more than 99%. The ablation study supports the claim that each component contributes, with the largest single gains coming from the global contrastive loss and the Laplace weighting at low label rates.

Load-bearing premise

The results rest on comparing each client's personalized local model, fine-tuned on its own labeled data, against baselines measured by their single shared global model; if that asymmetry is removed, the reported gains may shrink.

Editorial extensions

If this is right

  • At a 10% label rate, SSFL-DCSL reportedly matches or exceeds supervised federated training with 20% labels, so a factory could roughly halve its labeling effort for the same diagnostic accuracy.
  • Because clients exchange only prototype vectors (~0.06 MB per round) instead of full model weights (~8.6 MB), per-round communication drops by more than 99%, which matters for bandwidth-limited industrial links.
  • The method tolerates client dropout: losing one client's prototype lowers accuracy from 96.71% to 96.09% at a 40% label rate on the largest public benchmark, and losing three clients still leaves 93.57%.
  • The ablation path shows each proposed component contributes, with the global contrastive loss improving one benchmark by 13.03 points at 10% labels and TLAW adding 0.51–3.87 points, so the headline result is attributed to the full combination.

Reading between the lines

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

  • The reported comparison may not be apples-to-apples: the paper evaluates SSFL-DCSL as personalized models fine-tuned on each client's local labeled data, while FedAvg-style baselines are reported through their shared global model; giving baselines the same local fine-tuning could shrink the 1.15%–7.85% gap and would isolate the proposed components' true contribution.
  • The Laplace-distribution assumption on model confidence is testable: if the empirical histogram of $\max(p)$ over unlabeled batches deviates strongly from Laplace, the guaranteed lower bounds on pseudo-label quantity and quality in Theorem 1 may not hold in practice.
  • Prototype-only communication reduces bandwidth but changes the privacy surface: per-class mean embeddings can reveal which fault classes exist and roughly how clients cluster, so quantifying the information leakage and adding formal privacy protection is a natural extension.
  • Because the local contrastive loss selects positives and negatives using pseudo-labels, wrong pseudo-labels also corrupt pair selection; an extension would apply the same confidence-based weighting when sampling pairs, not only when weighting the unsupervised loss.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 4 minor

Summary. The manuscript proposes a semi-supervised federated learning framework, SSFL-DCSL, for intelligent fault diagnosis. The method combines a truncated-Laplace adaptive sample weighting function (TLAW) to down-weight low-confidence pseudo-labels, a dual contrastive loss made of a local contrastive loss (LCL) and a global contrastive loss (GCL), and prototype-based aggregation with momentum updates. Experiments are reported on the PU, MFPT, CWRU, and a factory chemical-pump dataset at 10%, 20%, and 40% label rates, including ablations, computational efficiency measurements, and a client-dropout study. The headline claim is a 1.15% to 7.85% accuracy improvement over state-of-the-art methods when only 10% of the data are labeled.

Significance. The problem addressed is relevant: label-scarce and heterogeneous federated fault diagnosis is a realistic industrial setting, and the prototype-only communication idea is attractive for privacy and bandwidth. If the claimed accuracy gains were established and the comparison were fair, SSFL-DCSL would be a meaningful contribution. The paper also provides a stepwise ablation study in Table V and evaluates on multiple datasets. However, the central empirical claim is not reproducible from the reported tables, and the experimental comparison appears to mix personalized local models with global-model baselines. In addition, the theoretical support for TLAW is not rigorously established. Until these load-bearing issues are resolved, the contribution cannot be assessed on the evidence presented.

major comments (4)
  1. [Abstract and Tables IV and VI] The abstract claims a 1.15% to 7.85% improvement over state-of-the-art methods at a 10% label rate, but this range is not recoverable from the reported results. At 10% label rate, Table IV gives best-baseline gaps of 2.64 points on PU (93.23 vs. FedCon at 90.59), 3.06 on MFPT (93.05 vs. FedCD at 89.99), and 2.31 on CWRU (91.50 vs. FedCD at 89.19); Table VI gives 6.58 on CP (86.01 vs. FedCD at 79.43). The observed range is therefore 2.31 to 6.58, not 1.15 to 7.85. Please state exactly which comparisons produce the reported endpoints.
  2. [Section III-A, Section IV-A4, and Algorithm 1] The evaluation does not isolate the proposed algorithmic components from personalization. Section III-A states that during the fine-tuning stage each client fine-tunes the model using its locally labeled data, and Section IV-A4 says results are averaged over clients, but no analogous fine-tuning is reported for FedAvg-Supervised, FedAvg/FedProx-FixMatch/UDA, Fed-SSMPN, FedCon, or FedCD. Moreover, Algorithm 1 only exchanges prototypes (lines 9-10) and never updates a global model, so SSFL-DCSL as implemented is effectively a personalized local-model method. The reported gains could therefore stem from per-client adaptation rather than from TLAW, LCL/GCL, or PTA. A controlled comparison with the same fine-tuning protocol applied to all methods, or to none, is needed.
  3. [Appendix A, proof of Theorem 1] The proof of Theorem 1 assumes 'without loss of generality' that exactly half of the samples satisfy max(p_i) < mu_t because mu_t is the sample mean of max(p) and hence 'P(max(p) < mu_t) = 0.5'. This conflates the sample mean with the median; for a finite sample, it is not true that exactly half of the observations fall below the mean. Equations (23)-(27) rely on splitting the sum at U/2, so the stated lower bounds on f(p) and g(p) are not established. A corrected proof, or an explicitly justified distributional assumption, is required before Theorem 1 can be used to support TLAW.
  4. [Section II and Algorithm 1] There is an internal inconsistency in the description of the federated algorithm. Section II specifies a standard FL loop in which local models are initialized by the global model and aggregated as theta_G <- sum_k (N_k/N) theta_k, but Algorithm 1 never performs this update: lines 9-10 only update global prototypes and local prototype sets, and Section III-C also describes only prototype communication. The role of the global model theta_G is therefore unclear. The authors should state clearly whether SSFL-DCSL is a personalized prototype-sharing method and, if so, rewrite the system model accordingly.
minor comments (4)
  1. [Abstract vs. Section IV-A1] The abstract says experiments are conducted on two publicly available datasets, while Section IV-A1 describes three openly accessible datasets (PU, MFPT, CWRU) plus the factory CP dataset; please correct the count.
  2. [Section IV-E] The client-dropout study in Table VII reports only the proposed method without a baseline subjected to the same dropout pattern, so the claim of high tolerance to stragglers is not comparative.
  3. [Notation, Eqs. (10)-(11) and (16)] The symbol N_i is used both for the negative-pair set in Eq. (10) and for the total number of class-j samples in Eq. (16), and P_i/P_j denote both positive-pair sets and prototypes; please use distinct symbols.
  4. [References] References [27] and [30] appear to be the same paper (same title, same volume, same pages, same authors) with different issue months; please deduplicate.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the headline accuracy claim is an empirical comparison against external benchmarks, and the TLAW bound is a derived inequality, not a fitted input renamed as a prediction.

full rationale

The paper's central claim is the reported 1.15–7.85% accuracy improvement over state-of-the-art methods at a 10% label rate (Abstract; Tables IV and VI). This is an empirical benchmark comparison against external baselines (FedAvg, FixMatch, UDA, FedCon, FedCD), not a quantity derived from the paper's own definitions or fitted parameters. The TLAW function (Eq. 4) estimates its Laplace parameters from the model's own confidence statistics via EMA (Eqs. 5–6); this is a disclosed self-referential SSL mechanism, but the paper does not present the resulting accuracy as a prediction from those fitted parameters. Theorem 1's quantity/quality lower bounds (Appendix A) follow from the definitions in Eqs. 7–8 and the explicit piecewise form of λ(p); they are mathematical inequalities about the weighting scheme, not a claim equivalent to its input. The fine-tuning stage (Section III-A) and prototype-only aggregation (Algorithm 1) raise a legitimate evaluation-protocol question about whether the comparison isolates personalization from the proposed losses, but that is an experimental attribution concern, not a derivation-chain circularity: no equation in the paper forces the reported gains by construction. Self-citations such as [13] and [50] appear only in related-work survey positions and are not load-bearing for the method's correctness or the empirical result.

Assumptions & free parameters 6 free parameters · 3 assumptions · 0 invented entities

The central claim rests on several hand-chosen hyperparameters (lambda_max, tau, alpha, m, kappa) and on two distributional assumptions about the model's confidence, one of which (sample mean equals median) is used in the proof but is not guaranteed. No new physical or conceptual entities are introduced.

free parameters (6)
  • lambda_max = not specified
    Maximum sample weight in the TLAW function (Eq. 4); a hand-chosen constant controlling pseudo-label weighting magnitude.
  • base contrastive temperature tau = not specified
    Base temperature in the dynamic temperature formula tau_i = tau * (1 + alpha * sigma_t) used in LCL and GCL.
  • temperature scaling alpha = not specified
    Scaling factor for the dynamic temperature; no value is given in the paper.
  • EMA momentum m = not specified
    Momentum for exponential moving average estimation of mu_t and sigma_t in Eq. (6).
  • prototype momentum kappa = not specified
    Momentum coefficient for global prototype updates in Eq. (18); no value is specified.
  • Dirichlet concentration nu = 0.5
    Controls the non-IID data partition across clients; chosen without sensitivity analysis.
assumptions (3)
  • domain assumption The confidence max(p) of model predictions follows a truncated Laplace distribution with mean mu_t and scale b_t.
    Stated in Section III-B2 before Eq. (4). This justifies the TLAW weighting but is not empirically validated.
  • ad hoc to paper For a finite sample, the mean mu_t of max(p) equals the median, so exactly half the samples lie below mu_t.
    Used in Appendix A, Eqs. (23)-(27), to derive quantity and quality bounds. This is not true for general finite samples and only holds asymptotically for symmetric distributions.
  • domain assumption Pseudo-labels generated by the model can serve as training targets for unlabeled data.
    Standard pseudo-labeling assumption invoked throughout Section III-B2; correctness depends on model confidence and the weighting mechanism.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Semi-Supervised Federated Learning via Dual Contrastive Learning and Soft Labeling for Intelligent Fault Diagnosis." pith.science (2026). https://pith.science/paper/DKJ63YG7

@misc{pith2026250714181,
  author       = {Pith},
  title        = {Pith review of: Semi-Supervised Federated Learning via Dual Contrastive Learning and Soft Labeling for Intelligent Fault Diagnosis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DKJ63YG7}},
  note         = {Machine review of arXiv:2507.14181}
}
read the original abstract

Intelligent fault diagnosis (IFD) plays a crucial role in ensuring the safe operation of industrial machinery and improving production efficiency. However, traditional supervised deep learning methods require a large amount of training data and labels, which are often located in different clients. Additionally, the cost of data labeling is high, making labels difficult to acquire. Meanwhile, differences in data distribution among clients may also hinder the model's performance. To tackle these challenges, this paper proposes a semi-supervised federated learning framework, SSFL-DCSL, which integrates dual contrastive loss and soft labeling to address data and label scarcity for distributed clients with few labeled samples while safeguarding user privacy. It enables representation learning using unlabeled data on the client side and facilitates joint learning among clients through prototypes, thereby achieving mutual knowledge sharing and preventing local model divergence. Specifically, first, a sample weighting function based on the Laplace distribution is designed to alleviate bias caused by low confidence in pseudo labels during the semi-supervised training process. Second, a dual contrastive loss is introduced to mitigate model divergence caused by different data distributions, comprising local contrastive loss and global contrastive loss. Third, local prototypes are aggregated on the server with weighted averaging and updated with momentum to share knowledge among clients. To evaluate the proposed SSFL-DCSL framework, experiments are conducted on two publicly available datasets and a dataset collected on motors from the factory. In the most challenging task, where only 10\% of the data are labeled, the proposed SSFL-DCSL can improve accuracy by 1.15% to 7.85% over state-of-the-art methods.

Figures

Figures reproduced from arXiv: 2507.14181 by the authors.

Figure 1
Figure 1. The overall framework of the local training stage of SSFL-DCSL. [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. The machines in our partner chemical factory. (a) [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Data distribution of four datasets for different clients. [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Diagram of the local user’s model: CNN and Trans [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Radar chart of proposed SSFL-DCSL computational [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

52 extracted references · 52 canonical work pages

  1. [1]

    Knowledge- based fault diagnosis in industrial internet of things: A survey,

    Y . Chi, Y . Dong, Z. J. Wang, F. R. Yu, and V . C. M. Leung, “Knowledge- based fault diagnosis in industrial internet of things: A survey,” IEEE Internet Things J. , vol. 9, no. 15, pp. 12 886–12 900, Mar., 2022

    work page 2022

  2. [2]

    Fault diagnosis of rotating machinery based on 1d–2d joint convolution neural network,

    W. Du, P. Hu, H. Wang, X. Gong, and S. Wang, “Fault diagnosis of rotating machinery based on 1d–2d joint convolution neural network,” IEEE Trans. Ind. Electron. , vol. 70, no. 5, pp. 5277–5285, Jun., 2023

    work page 2023

  3. [3]

    Self-training reinforced adversarial adaptation for machine fault diagnosis,

    J. Jiao, H. Li, and J. Lin, “Self-training reinforced adversarial adaptation for machine fault diagnosis,” IEEE Trans. Ind. Electron., vol. 70, no. 11, pp. 11 649–11 658, Dec., 2023

    work page 2023

  4. [4]

    Adaptive fault components extraction by using an optimized weights spectrum based index for machinery fault diagnosis,

    B. Hou, D. Wang, Z. Peng, and K.-L. Tsui, “Adaptive fault components extraction by using an optimized weights spectrum based index for machinery fault diagnosis,” IEEE Trans. Ind. Electron. , vol. 71, no. 1, pp. 985–995, Feb., 2024

    work page 2024

  5. [5]

    A novel generative adver- sarial networks via music theory knowledge for early fault intelligent diagnosis of motor bearings,

    P. Luo, Z. Yin, D. Yuan, F. Gao, and J. Liu, “A novel generative adver- sarial networks via music theory knowledge for early fault intelligent diagnosis of motor bearings,” IEEE Trans. Ind. Electron., vol. 71, no. 8, pp. 9777–9788, Oct., 2024

    work page 2024

  6. [6]

    Label recovery and trajectory designable network for transfer fault diagnosis of machines with incorrect annotation,

    B. Yang, Y . Lei, X. Li, N. Li, and A. K. Nandi, “Label recovery and trajectory designable network for transfer fault diagnosis of machines with incorrect annotation,” IEEE CAA J. Autom. Sinica , vol. 11, no. 4, pp. 932–945, Apr., 2024

    work page 2024

  7. [7]

    Uncertainty-aware deep learning: A promising tool for trustworthy fault diagnosis,

    J. Ren, J. Wen, Z. Zhao, R. Yan, X. Chen, and A. K. Nandi, “Uncertainty-aware deep learning: A promising tool for trustworthy fault diagnosis,” IEEE CAA J. Autom. Sinica , vol. 11, no. 6, pp. 1317–1330, May., 2024

    work page 2024

  8. [8]

    Communication-efficient learning of deep networks from decentralized data,

    B. McMahan, E. Moore, D. Ramage, S. Hampson, and B. A. y Arcas, “Communication-efficient learning of deep networks from decentralized data,” in Int. Conf. Artif. Intell. Stat., vol. 54, Apr., 2017, pp. 1273–1282

    work page 2017

Show all 52 references
  1. [9]

    Intelligent fault diagnosis of rolling bearings with low-quality data: A feature significance and diversity learning method,

    J. Chen, C. Lin, B. Yao, L. Yang, and H. Ge, “Intelligent fault diagnosis of rolling bearings with low-quality data: A feature significance and diversity learning method,” Reliab. Eng. Syst. Saf. , vol. 237, p. 109343, Sep., 2023

    2023

  2. [10]

    Federated learning in intelligent transportation systems: Recent applications and open problems,

    S. Zhang, J. Li, L. Shi, M. Ding, D. C. Nguyen, W. Tan, J. Weng, and Z. Han, “Federated learning in intelligent transportation systems: Recent applications and open problems,” IEEE Trans. Intell. Transp. Syst., vol. 25, no. 5, pp. 3259–3285, May., 2024

    2024

  3. [11]

    An effective federated learning verification strategy and its applications for fault diagnosis in industrial iot systems,

    Y . Li, Y . Chen, K. Zhu, C. Bai, and J. Zhang, “An effective federated learning verification strategy and its applications for fault diagnosis in industrial iot systems,” IEEE Internet Things J. , vol. 9, no. 18, pp. 16 835–16 849, Feb., 2022

    2022

  4. [12]

    Bearing fault diagnosis based on improved federated learning algorithm,

    D. Geng, H. He, X. Lan, and C. Liu, “Bearing fault diagnosis based on improved federated learning algorithm,” Computing, vol. 104, no. 1, pp. 1–19, Oct., 2022

    2022

  5. [13]

    Blockchain assisted decentralized federated learning (BLADE- FL): performance analysis and resource allocation,

    J. Li, Y . Shao, K. Wei, M. Ding, C. Ma, L. Shi, Z. Han, and H. V . Poor, “Blockchain assisted decentralized federated learning (BLADE- FL): performance analysis and resource allocation,”IEEE Trans. Parallel Distrib. Syst., vol. 33, no. 10, pp. 2401–2415, Dec., 2022

    2022

  6. [14]

    Lightweight fl: A low- cost federated learning framework for mechanical fault diagnosis with training optimization and model pruning,

    J. Du, N. Qin, D. Huang, X. Jia, and Y . Zhang, “Lightweight fl: A low- cost federated learning framework for mechanical fault diagnosis with training optimization and model pruning,” IEEE Trans. Instrum. Meas. , vol. 73, pp. 1–14, 2024

    2024

  7. [15]

    Trusted AI in multi-agent systems: An overview of privacy and security for distributed learning,

    C. Ma, J. Li, K. Wei, B. Liu, M. Ding, L. Yuan, Z. Han, and H. V . Poor, “Trusted AI in multi-agent systems: An overview of privacy and security for distributed learning,” CoRR, vol. abs/2202.09027, Aug., 2022

    2022 arXiv

  8. [16]

    Fedalign: Federated model alignment via data-free knowledge distillation for machine fault diagnosis,

    W. Sun, R. Yan, R. Jin, R. Zhao, and Z. Chen, “Fedalign: Federated model alignment via data-free knowledge distillation for machine fault diagnosis,” IEEE Trans. Instrum. Meas. , vol. 73, pp. 1–12, 2024

    2024

  9. [17]

    Adaptive privacy-preserving federated learning for fault diagnosis in internet of ships,

    Z. Zhang, C. Guan, H. Chen, X. Yang, W. Gong, and A. Yang, “Adaptive privacy-preserving federated learning for fault diagnosis in internet of ships,” IEEE Internet Things J., vol. 9, no. 9, pp. 6844–6854, Sep., 2022

    2022

  10. [18]

    Fedcae: A new federated learning framework for edge-cloud collaboration based machine fault diagnosis,

    Y . Yu, L. Guo, H. Gao, Y . He, Z. You, and A. Duan, “Fedcae: A new federated learning framework for edge-cloud collaboration based machine fault diagnosis,” IEEE Trans. Ind. Electron., vol. 71, no. 4, pp. 4108–4119, May., 2024

    2024

  11. [19]

    Domain adversarial transfer network for cross-domain fault diagnosis of rotary machinery,

    Z. Chen, G. He, J. Li, Y . Liao, K. C. Gryllias, and W. Li, “Domain adversarial transfer network for cross-domain fault diagnosis of rotary machinery,”IEEE Trans. Instrum. Meas., vol. 69, no. 11, pp. 8702–8712, May., 2020

    2020

  12. [21]

    Federated generalized zero-sample industrial fault diagnosis across multisource domains,

    L. Cai, H. Yin, J. Lin, and Y . Hu, “Federated generalized zero-sample industrial fault diagnosis across multisource domains,” IEEE Internet Things J., vol. 11, no. 23, pp. 38 895–38 906, Sep., 2024

    2024

  13. [22]

    Intelligent fault diagnosis via ring-based decentralized federated transfer learning,

    L. Wan, J. Ning, Y . Li, C. Li, and K. Li, “Intelligent fault diagnosis via ring-based decentralized federated transfer learning,” Knowl. Based Syst., vol. 284, p. 111288, 2024

    2024

  14. [23]

    Industrial edge intelligence: Federated-meta learning framework for few-shot fault diagnosis,

    J. Chen, J. Tang, and W. Li, “Industrial edge intelligence: Federated-meta learning framework for few-shot fault diagnosis,” IEEE Trans. Netw. Sci. Eng., vol. 10, no. 6, pp. 3561–3573, Nov., 2023

    2023

  15. [24]

    In defense of pseudo-labeling: An uncertainty-aware pseudo-label selection frame- work for semi-supervised learning,

    M. N. Rizve, K. Duarte, Y . S. Rawat, and M. Shah, “In defense of pseudo-labeling: An uncertainty-aware pseudo-label selection frame- work for semi-supervised learning,” in Int. Conf. Learn. Represent. , May., 2021

    2021

  16. [25]

    Regularization with stochastic transformations and perturbations for deep semi-supervised learning,

    M. Sajjadi, M. Javanmardi, and T. Tasdizen, “Regularization with stochastic transformations and perturbations for deep semi-supervised learning,” in Adv. Neural Inf. Process. Syst., Dec., 2016, pp. 1163–1171

    2016

  17. [26]

    Semisupervised momentum prototype network for gearbox fault diagnosis under limited 13 labeled samples,

    X. Zhang, Z. Su, X. Hu, Y . Han, and S. Wang, “Semisupervised momentum prototype network for gearbox fault diagnosis under limited 13 labeled samples,” IEEE Trans. Ind. Inf. , vol. 18, no. 9, pp. 6203–6213, Sep., 2022

    2022

  18. [27]

    Online fault diagnosis of harmonic drives using semisupervised contrastive graph generative network via multimodal data,

    G. Yang, H. Tao, T. Yu, R. Du, and Y . Zhong, “Online fault diagnosis of harmonic drives using semisupervised contrastive graph generative network via multimodal data,” IEEE Trans. Ind. Electron., vol. 71, no. 3, pp. 3055–3063, Mar., 2024

    2024

  19. [28]

    Semisupervised fault diagnosis of gearbox using weighted graph-based label propagation and virtual adversarial training,

    D. Shan, C. Cheng, L. Li, Z. Peng, and Q. He, “Semisupervised fault diagnosis of gearbox using weighted graph-based label propagation and virtual adversarial training,” IEEE Trans. Instrum. Meas. , vol. 72, pp. 1–11, Nov., 2023

    2023

  20. [29]

    Meta-self-training based on teacher–student network for industrial label-noise fault diagnosis,

    X. Pu and C. Li, “Meta-self-training based on teacher–student network for industrial label-noise fault diagnosis,” IEEE Trans. Instrum. Meas. , vol. 72, pp. 1–11, Dec., 2023

    2023

  21. [30]

    Online fault diagnosis of harmonic drives using semisupervised contrastive graph generative network via multimodal data,

    G. Yang, H. Tao, T. Yu, R. Du, and Y . Zhong, “Online fault diagnosis of harmonic drives using semisupervised contrastive graph generative network via multimodal data,” IEEE Trans. Ind. Electron., vol. 71, no. 3, pp. 3055–3063, Apr., 2024

    2024

  22. [31]

    Fixmatch: Simplifying semi- supervised learning with consistency and confidence,

    K. Sohn, D. Berthelot, C.-L. Li, Z. Zhang, N. Carlini, E. D. Cubuk, A. Kurakin, H. Zhang, and C. Raffel, “Fixmatch: Simplifying semi- supervised learning with consistency and confidence,” in Adv. Neural Inf. Process. Syst. , vol. 33, Dec., 2020, pp. 596–608

    2020

  23. [32]

    Federated optimization in heterogeneous networks,

    T. Li, A. K. Sahu, M. Zaheer, M. Sanjabi, A. Talwalkar, and V . Smith, “Federated optimization in heterogeneous networks,” in Proc. Mach. Learn. Syst., Mar., 2020

    2020

  24. [33]

    Fedcon: A con- trastive framework for federated semi-supervised learning,

    Z. Long, J. Wang, Y . Wang, H. Xiao, and F. Ma, “Fedcon: A con- trastive framework for federated semi-supervised learning,” CoRR, vol. abs/2109.04533, Sep., 2021

    2021 arXiv

  25. [34]

    Fedcd: Federated semi-supervised learning with class awareness balance via dual teachers,

    Y . Liu, H. Wu, and J. Qin, “Fedcd: Federated semi-supervised learning with class awareness balance via dual teachers,” in Proc. AAAI Conf. Artif. Intell., vol. 38, no. 4, Mar., 2024, pp. 3837–3845

    2024

  26. [35]

    Rscfed: Random sampling consensus federated semi-supervised learning,

    X. Liang, Y . Lin, H. Fu, L. Zhu, and X. Li, “Rscfed: Random sampling consensus federated semi-supervised learning,” in Proc. IEEE Comput. Soc. Conf. Comput. Vis. Pattern Recognit. , Mar., 2022, pp. 10 144– 10 153

    2022

  27. [36]

    Estimating before debiasing: A bayesian approach to detaching prior bias in federated semi-supervised learning,

    G. Zhu, X. Liu, X. Wu, S. Tang, C. Tang, J. Niu, and H. Su, “Estimating before debiasing: A bayesian approach to detaching prior bias in federated semi-supervised learning,” in Int. Jt. Conf. Artif. Intell. ijcai.org, May., 2024, pp. 2625–2633

    2024

  28. [37]

    (fl) 2: Overcoming few labels in federated semi-supervised learning,

    S. Lee, T. V . Le, J. Shin, and S. Lee, “(fl) 2: Overcoming few labels in federated semi-supervised learning,” in Proc. Annu. Conf. Neural Inf. Process. Syst. , A. Globersons, L. Mackey, D. Belgrave, A. Fan, U. Paquet, J. M. Tomczak, and C. Zhang, Eds., Oct., 2024

    2024

  29. [38]

    Federated transfer learning for bearing fault diagnosis with discrepancy-based weighted federated averaging,

    J. Chen, J. Li, R. Huang, K. Yue, Z. Chen, and W. Li, “Federated transfer learning for bearing fault diagnosis with discrepancy-based weighted federated averaging,” IEEE Trans. Instrum. Meas. , vol. 71, pp. 1–11, Jun., 2022

    2022

  30. [39]

    Federated distillation with lightweight generative adversarial network for servo motor bearing fault diagnosis in heterogeneous data,

    Z. Liu, Q. Hao, H. Shao, R. Xin, and K. Zhao, “Federated distillation with lightweight generative adversarial network for servo motor bearing fault diagnosis in heterogeneous data,” IEEE Internet Things J. , vol. 12, no. 9, pp. 11 744–11 753, Feb., 2025

    2025

  31. [40]

    A gradient alignment federated do- main generalization framework for rotating machinery fault diagnosis,

    Z. Ye, J. Wu, X. He, and W. Jiang, “A gradient alignment federated do- main generalization framework for rotating machinery fault diagnosis,” IEEE Internet Things J. , pp. 1–1, Apr., 2025

    2025

  32. [41]

    Heterogeneous federated learning: Client-side collaborative update interdomain gener- alization method for intelligent fault diagnosis,

    H. Ma, J. Wei, G. Zhang, Q. Wang, X. Kong, and J. Du, “Heterogeneous federated learning: Client-side collaborative update interdomain gener- alization method for intelligent fault diagnosis,” IEEE Internet Things J., vol. 12, no. 5, pp. 5704–5718, Mar., 2025

    2025

  33. [42]

    Unsupervised data augmentation for consistency training,

    Q. Xie, Z. Dai, E. H. Hovy, T. Luong, and Q. Le, “Unsupervised data augmentation for consistency training,” inAdv. Neural Inf. Process. Syst., Dec., 2020

    2020

  34. [43]

    A systematic survey on federated semi-supervised learning,

    Z. Song, X. Yang, Y . Zhang, X. Fu, Z. Xu, and I. King, “A systematic survey on federated semi-supervised learning,” in Int. Jt. Conf. Artif. Intell. ijcai.org, Aug., 2024, pp. 8244–8252

    2024

  35. [44]

    Semi- supervised multiscale permutation entropy-enhanced contrastive learning for fault diagnosis of rotating machinery,

    Y . Zhou, H. Wang, G. Wang, A. Kumar, W. Sun, and J. Xiang, “Semi- supervised multiscale permutation entropy-enhanced contrastive learning for fault diagnosis of rotating machinery,” IEEE Trans. Instrum. Meas. , vol. 72, pp. 1–10, Aug., 2023

    2023

  36. [45]

    Semi- supervised contrastive domain adaptation network for fault diagnosis of rotating machinery under cross-working conditions,

    X. Lu, L. Song, C. Han, Q. Jiang, W. Xu, and H. Wang, “Semi- supervised contrastive domain adaptation network for fault diagnosis of rotating machinery under cross-working conditions,” IEEE Internet Things J., pp. 1–1, Mar., 2025

    2025

  37. [46]

    Drsc: Dual-reweighted siamese contrastive learning network for cross-domain rotating machinery fault diagnosis with multi-source domain imbalanced data,

    Y . Bi, R. Fu, C. Jiang, X. Zhang, F. Li, L. Zhao, and G. Han, “Drsc: Dual-reweighted siamese contrastive learning network for cross-domain rotating machinery fault diagnosis with multi-source domain imbalanced data,” IEEE Internet Things J. , pp. 1–1, Mar., 2025

    2025

  38. [47]

    Open-set fault diagnosis via supervised contrastive learning with negative out- of-distribution data augmentation,

    P. Peng, J. Lu, T. Xie, S. Tao, H. Wang, and H. Zhang, “Open-set fault diagnosis via supervised contrastive learning with negative out- of-distribution data augmentation,” IEEE Trans. Ind. Informat. , vol. 19, Mar., 2023

    2023

  39. [48]

    Class incremental fault diagnosis under limited fault data via supervised contrastive knowledge distillation,

    H. Zhang, Y . Yao, Z. Wang, J. Su, M. Li, P. Peng, and H. Wang, “Class incremental fault diagnosis under limited fault data via supervised contrastive knowledge distillation,” IEEE Trans. Ind. Informat. , pp. 1– 11, Mar., 2025

    2025

  40. [49]

    Domain discrepancy-guided contrastive feature learning for few-shot industrial fault diagnosis under variable working conditions,

    T. Zhang, J. Chen, S. Liu, and Z. Liu, “Domain discrepancy-guided contrastive feature learning for few-shot industrial fault diagnosis under variable working conditions,”IEEE Trans. Ind. Informat., vol. 19, no. 10, pp. 10 277–10 287, Oct., 2023

    2023

  41. [50]

    Self-supervised learning for multimodal fault diagnosis with shapley-value weighted transformers,

    Y . Dai, J. Li, Z. Mei, Y . Ni, S. Guo, and Z. Li, “Self-supervised learning for multimodal fault diagnosis with shapley-value weighted transformers,” IEEE Trans. Instrum. Meas. , vol. 74, pp. 1–14, May., 2025

    2025

  42. [51]

    A contrastive learning-based fault diagnosis method for rotating machinery with limited and imbalanced labels,

    Y . Zhang, Z. Liu, and Q. Huang, “A contrastive learning-based fault diagnosis method for rotating machinery with limited and imbalanced labels,” IEEE Sens. J. , vol. 23, no. 14, pp. 16 402–16 412, Jul., 2023

    2023

  43. [52]

    Pseudo-label: The simple and efficient semi-supervised learning method for deep neural networks,

    D.-H. Lee et al., “Pseudo-label: The simple and efficient semi-supervised learning method for deep neural networks,” in Proc. Int. Conf. Mach. Learn., vol. 3, no. 2, 2013, p. 896

    2013

  44. [53]

    Time-series representation learning via temporal and contextual con- trasting,

    E. Eldele, M. Ragab, Z. Chen, M. Wu, C. K. Kwoh, X. Li, and C. Guan, “Time-series representation learning via temporal and contextual con- trasting,” in Int. Jt. Conf. Artif. Intell. , Aug., 2021, pp. 2352–2359. Yajiao Dai received a B.S. degree in Biomedi- cal Engineering fro...

    2021

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.