Pith. sign in

REVIEW 4 major objections 6 minor 41 references

Learning Compositional Transferability of Time Series for Source-Free Domain Adaptation

T0 review · 4 major / 6 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read Source-free time series domain adaptation can be composed from a frozen source-trained U-net replay and a small learnable warp autoencoder, with test-time per-instance rescaling.

desk verdict A clever compositional reconstruction method for SFDA whose empirical claim is undercut by a pre-training protocol difference versus baselines; worth a round of revision, not a desk reject. read the letter →

arxiv 2504.14994 v1 pith:7ZXI4GEZ submitted 2025-04-21 cs.LG

Hankang Sun (1) , Guiming Li (1) , Su Yang (1) , Baoqi Li (2) ((1) Fudan University , (2) Chinese Academy of Sciences , Institute of Acoustics) This is my paper · ORCID
classification cs.LG
keywords source-freedomainadaptationtimeseriesclassificationcompositionalreconstructionU-netwarpautoencodertest-timeTsallisentropyinstance-wiserescaling
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 solve source-free domain adaptation for time series classification, the setting where neither source data nor target labels are available during adaptation. Its central idea is that a single fine-tuned reconstructor squanders the source prior and cannot accommodate heterogeneous temporal patterns, so the transfer should be composed from two stages: a frozen source-trained U-net replays coarse source knowledge, and a small learnable autoencoder compensates the remaining offset. The two outputs are combined with learnable scaling factors, and at test time the source-replay factor is perturbed per instance and classifier outputs are ensembled by a cosine-similarity stability weight. The paper reports state-of-the-art macro-F1 scores on MFD, SSC, and UCIHAR, improving on prior source-free methods by 3.7%, 0.78%, and 2.6% with instance rescaling. The result matters because it promises a parameter-light, backbone-agnostic way to reuse source reconstruction knowledge without any source data.

What carries the argument

The central object is the compositional two-branch reconstruction module: a frozen source-pretrained U-net provides source replay, and its output feeds two parallel branches, a residual source-replay branch that preserves the prior and an offset-compensation branch built from a VQ-VAE-style warp block (a vector-quantised variational autoencoder variant). A learnable scaling factor $v_T$ weights the warp branch while the replay branch keeps factor $v_S=1$ during group adaptation, forming the final reconstruction in Eq. (2) as $\hat{X}^T_i = v_T h_\phi(j_\theta(X^T_i)) + v_S j_\theta(X^T_i)$. This mechanism carries the argument because optimizing the target reconstruction objective $\mathcal{L}^T_{\mathrm{MSE}}$ (Eq. 3) plus improved Tsallis entropy loss is what aligns target series to the source distribution, with the frozen U-net supplying the bulk of the transfer and the small warp block supplying the residual correction. The test-time instance-wise adaptation perturbs the source-replay scaling factor and ensembles classifier outputs weighted by cosine-similarity stability.

What would settle it

To test this, one could construct a controlled source/target pair where the compositional module drives target reconstruction MSE to near zero while the frozen classifier's accuracy on the reconstructed target series remains at chance level; if such a pair appears on MFD, SSC, or UCIHAR, the alignment assumption fails.

Watch

Extended reading notes

Core claim

The paper's central claim is that domain transferability for time series can be disentangled into a coarse source-replay step and a fine offset-compensation step, and that this decomposition outperforms fine-tuning a single reconstruction model under source-free constraints. Concretely, target series are first run through a frozen U-net reconstructor pre-trained on source data; the output is carried forward by a residual source-replay branch (scaling factor fixed at 1 during group adaptation) and simultaneously warped by a small VQ-VAE-style autoencoder with a learnable scaling factor. The final reconstructed series is the weighted sum of the two branches, optimized on unlabeled target data by reconstruction MSE plus improved Tsallis entropy. At inference, the source-replay scaling factor is swept within a small neighborhood, cosine similarity between successive classifier outputs measures stability, and softmax-weighted ensembling of the classifier outputs yields the instance-adaptive prediction. The paper states that this compositional architecture achieves state-of-the-art MF1 scores of 96.15%, 64.83%, and 92.17% on the three benchmarks.

Load-bearing premise

The method assumes that minimizing reconstruction error on unlabeled target data is enough to make the reconstructed target series land in classes the frozen source classifier recognizes; if reconstruction fidelity does not imply class discriminability, the adaptation can lower reconstruction loss without improving classification.

Editorial extensions

If this is right

  • The source-replay branch alone accounts for the majority of the domain-gap reduction, with offset compensation adding a finer correction (Figure 6), so the frozen U-net's prior is the main engine of transfer.
  • Because adaptation operates on reconstructed time series rather than latent features, the method can be attached to any frozen classification backbone without knowing its internal structure.
  • Perturbing the source-replay scaling factor at test time and ensembling by cosine-similarity stability improves over group-level adaptation alone, giving a cheap instance-wise adaptation mechanism.
  • The learnable part of the adapter is tiny (8,170 parameters for the warp block against 34.53 million frozen in the U-net), so the method is computationally light enough for deployment.
  • Removing either branch degrades performance substantially (e.g., MFD drops from 96.15% to 73.26% without the source-replay branch), so both the prior-preserving replay and the offset compensation are load-bearing.

Reading between the lines

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

  • This suggests a general recipe: freeze a high-capacity source prior and learn only a small residual corrector, which might extend to other source-free reconstruction tasks beyond time series, such as images or audio, where a similar coarse-to-fine split could be tested.
  • The paper's alignment premise is that reconstruction error is a faithful proxy for classification alignment; a natural stress test would corrupt target class priors while keeping reconstruction easy, which would reveal whether the proxy breaks.
  • The cosine-similarity stability ensemble is a form of perturbation-based uncertainty; combining it with other input perturbations, like noise on the reconstructed series, could yield further gains.
  • The improvement varies strongly across benchmarks (3.7%, 0.78%, 2.6%), suggesting the method may pay off most when source and target share strong structural priors; applying it to datasets with larger domain gaps would clarify when the compositional split is sufficient.
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 / 6 minor

Summary. The paper proposes CT-SFDA, a source-free domain adaptation method for time series classification. A U-net reconstructor is pre-trained on source data and frozen; for target adaptation, the U-net output serves as a coarse source replay, and a lightweight autoencoder-based warp block with a learned scaling factor provides offset compensation. Group-level adaptation minimizes target reconstruction MSE plus a Tsallis entropy term, and during inference the source replay branch is perturbed by a small scaling factor while classifier outputs are ensembled with cosine-similarity-based stability weights. The authors report state-of-the-art MF1 scores on the MFD, SSC, and UCIHAR benchmarks, with improvements of 3.7%, 0.78%, and 2.6% over MAPU, and support the design with ablations of the two branches, the loss terms, and the instance-wise adaptation module.

Significance. If the empirical claims hold under a controlled comparison, the compositional reconstruction idea is a reasonable and parameter-efficient direction for source-free time series domain adaptation: it preserves the frozen source prior, adds only about 8k trainable parameters beyond the pre-trained reconstructor, and the ablation tables give initial evidence that both the source replay and offset compensation branches contribute. The qualitative visualizations in Figures 5 and 6 are consistent with the claimed coarse-to-fine transfer mechanism. However, the main SOTA claim is currently confounded by a source pre-training protocol mismatch and is not supported by uncertainty estimates, so the practical significance of the reported gains is not yet established.

major comments (4)
  1. [3.3, 4.2, Tables 4-6] The comparison with MAPU and the other baselines is not apples-to-apples. Section 3.3 states that the classification backbone is trained on the output of the frozen U-net reconstructor, and Section 4.2 confirms that 'the source time series reconstructed by the frozen U-net is used as the input to train the classification backbone.' In contrast, MAPU and the cited UDA baselines pre-train their classifiers on raw source signals under their own protocols. This changes the input distribution seen by the classifier and could alone account for part or all of the reported gains, since the U-net front-end may denoise or normalize the source data. The ablations in Tables 7 and 8 vary the reconstruction modules and branches but always keep the classifier trained on reconstructed source data, so they cannot isolate the contribution of the compositional adaptation. Please add a control experiment: (i) CT with the classifier trained on raw source data and (ii) MAPU (or a same-protocol baseline) with its classifier trained on reconstructed source data, using identical source pre-training and reporting per-scenario results.
  2. [Tables 4-6] The main results report a single MF1 value per scenario without standard deviations, number of runs, or significance tests. The average improvement over MAPU on SSC is only 0.78 percentage points, and Table 5 shows that CT is worse than MAPU on four of the five SSC tasks (16->1, 9->14, 12->5, 7->18), with the entire average gain coming from the 0->11 task (53.3 vs 43.36). With no repeated-run statistics, the claim of state-of-the-art performance on SSC is not established, and the aggregate improvements on the other datasets should also be assessed against run-to-run variation. Please report mean +/- standard deviation over at least three seeds per scenario and apply a paired significance test against MAPU.
  3. [3.4, 3.5, Figures 5-6] The core premise that minimizing target reconstruction error is a sufficient alignment signal for the frozen source classifier is not directly tested. The objective in Eq. (3) and Eq. (4) optimizes reconstruction MSE on target data, with no explicit coupling between reconstruction fidelity and class discriminability; the only supporting evidence is the qualitative t-SNE plot and the averaged nearest-neighbor distance in Figures 5 and 6. Please provide a quantitative analysis of the relationship between reconstruction loss and target classification MF1 during group-level adaptation, for example per-epoch curves or a correlation, or an ablation in which a same-capacity non-reconstructive warp is optimized with the same entropy objective. This would show that reconstruction, rather than the trainable front-end alone, drives the reported gains.
  4. [Eq. (5)-(7)] The test-time ensemble weights are based on the stability of the classifier's own outputs under small perturbations of v_S, which measures self-consistency rather than correctness. A confidently wrong but stable prediction would receive a high weight. The ablation in Table 10 shows only small gains from this module, so this is not the main weakness, but the authors should state this limitation explicitly and, if possible, compare against an oracle or a correctness-labeled stability measure on a subset of target data.
minor comments (6)
  1. [4.2] There is a typo in the Experimental Setup paragraph: 'thhe' should be 'the', and 'tabel' in Section 5.1 should be 'table'.
  2. [4.2, Table 3] The three training stages in Table 3 are described as 'training reconstructor, backbone network, and warp block in order,' but the warp block training is the group-level adaptation stage on target data; please clarify the terminology to distinguish source pre-training from target adaptation.
  3. [Table 7] The entries 'AE + AE' and 'AE + U-net' in Table 7 are not described in the text; please specify how the AE replaces the U-net as the source reconstructor, including its input representation and parameter count.
  4. [Eq. (5)-(7)] The index notation for the perturbation span is inconsistent: Eq. (5) defines j = 0, 1, ..., n, while Eq. (6) and Eq. (7) sum over k = -n to n and j = -n+1 to n. Please define the signed index range explicitly.
  5. [Figure 5 caption] The caption states that black lines indicate the course of migrating target data, but a static t-SNE scatter plot does not show trajectories; please explain how the migration paths are obtained or relabel the figure as a schematic.
  6. [Abstract] The abstract contains a stray footnote marker 'experimental results2' with no corresponding footnote; please remove it.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity found: the derivation is empirical, and the self-referential losses are standard unsupervised objectives rather than hidden uses of the target labels.

full rationale

The paper's central claim is an empirical benchmark result, not a theorem derived from its own definitions. The pre-training stage (Eq. 1) trains the U-net on source time series with MSE, and the classifier is then trained on the U-net reconstructions. The adaptation stage (Eqs. 2-4) trains only the warp block and the scaling factor v_T on unlabeled target data, using the target reconstruction error L_MSE plus an improved Tsallis entropy term L_UR. No equation is defined in terms of the evaluation labels or the final MF1 score; L_MSE uses the raw target time series as supervision, and L_UR is an unsupervised confidence term. The instance-wise stage (Eqs. 5-7) perturbs the fixed source-replay scaling factor and forms a convex combination of the classifier outputs weighted by cosine similarities between consecutive outputs; this is an ensemble/smoothing operation over the same unlabeled outputs, not a fit to target labels. The reported improvements over MAPU are comparisons on external benchmarks, and the ablations in Tables 7-10 test the components independently. The paper does cite prior work for the benchmark setting, the backbone structure, and the time-series-as-image idea, but those citations are external and not load-bearing self-citations, and no uniqueness theorem or ansatz is imported from the authors' own prior work. A possible confound is that the classifier is pre-trained on U-net-reconstructed source data rather than raw source data, which may make the comparison with MAPU unfair, but that is a correctness or experimental-design concern, not circularity: the adaptation objective does not reduce to the evaluation metric by construction. Therefore the appropriate circularity score is 0.

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

The method depends on a small set of numeric hyperparameters (lambda, Delta, n, schedules, reshape sizes) and several domain assumptions about the validity of reconstruction-based alignment. No new physical entities are introduced. The central contribution is architectural and empirical.

free parameters (6)
  • Tsallis entropy weight lambda = 0.1
    Set uniformly in Section 4.2; balances the unsupervised entropy term against reconstruction MSE in Eq. (4).
  • Perturbation step Delta for v_S = 0.1%
    Defined per dataset in Section 4.2; grid step for the test-time rescaling in Eq. (5).
  • Maximum span n for v_S search = 10 (MFD), 8 (SSC), 3 (UCIHAR)
    Per-dataset tuning of the test-time search range; a larger n broadens the ensemble.
  • Learning rates and epochs per stage = See Table 3, e.g., MFD: 5e-3/8, 2e-3/20, 5e-3/8
    Three training stages use dataset-specific schedules, making reported results conditional on these choices.
  • Scaling factor v_T = Learned during group adaptation
    Trained on target data in Eq. (2); sets the offset compensation's weight in the final reconstruction.
  • Reshape and zero-padding sizes = 1x64x80 (MFD), 1x48x64 (SSC), 9x64x64 (UCIHAR)
    Hand-chosen transforms of time series to image-like inputs; the paper states size does not affect performance without evidence.
assumptions (6)
  • domain assumption Source and target share the same label space with only feature distribution shift.
    Assumed in Section 3.1 following MAPU; if label spaces differ, the adaptation objective is undefined.
  • domain assumption Minimizing target reconstruction error of the two-branch model aligns target features with source for classification.
    Core alignment mechanism in Eq. (3); no proof or analysis that reconstruction fidelity implies class discriminability.
  • domain assumption The source pre-trained U-net has captured source distribution well enough to act as a replay prior.
    Invoked in Section 3.3; no source reconstruction quality metrics are reported.
  • ad hoc to paper Zero-padding time series into image-like grids is acceptable for U-net reconstruction.
    Section 4.2 pads to fixed sizes; the information loss and distortion of temporal structure from padding are not examined.
  • domain assumption Improved Tsallis entropy on classifier outputs improves target adaptation.
    Added in Eq. (4); Table 9 shows empirical gain, but no theoretical justification.
  • domain assumption Cosine similarity of classifier outputs under v_S perturbation is a valid stability measure for ensembling.
    Section 3.5 basis for weights in Eq. (6); the heuristic is not compared against other confidence measures beyond entropy.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Learning Compositional Transferability of Time Series for Source-Free Domain Adaptation." pith.science (2026). https://pith.science/paper/7ZXI4GEZ

@misc{pith2026250414994,
  author       = {Pith},
  title        = {Pith review of: Learning Compositional Transferability of Time Series for Source-Free Domain Adaptation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7ZXI4GEZ}},
  note         = {Machine review of arXiv:2504.14994}
}
read the original abstract

Domain adaptation is challenging for time series classification due to the highly dynamic nature. This study tackles the most difficult subtask when both target labels and source data are inaccessible, namely, source-free domain adaptation. To reuse the classification backbone pre-trained on source data, time series reconstruction is a sound solution that aligns target and source time series by minimizing the reconstruction errors of both. However, simply fine-tuning the source pre-trained reconstruction model on target data may lose the learnt priori, and it struggles to accommodate domain varying temporal patterns in a single encoder-decoder. Therefore, this paper tries to disentangle the composition of domain transferability by using a compositional architecture for time series reconstruction. Here, the preceding component is a U-net frozen since pre-trained, the output of which during adaptation is the initial reconstruction of a given target time series, acting as a coarse step to prompt the subsequent finer adaptation. The following pipeline for finer adaptation includes two parallel branches: The source replay branch using a residual link to preserve the output of U-net, and the offset compensation branch that applies an additional autoencoder (AE) to further warp U-net's output. By deploying a learnable factor on either branch to scale their composition in the final output of reconstruction, the data transferability is disentangled and the learnt reconstructive capability from source data is retained. During inference, aside from the batch-level optimization in the training, we search at test time stability-aware rescaling of source replay branch to tolerate instance-wise variation. The experimental results show that such compositional architecture of time series reconstruction leads to SOTA performance on 3 widely used benchmarks.

Figures

Figures reproduced from arXiv: 2504.14994 by the authors.

Figure 1
Figure 1. Composition of Domain Transferability (The latent features are illustrated in different [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Network architecture. 3 Compositional Time Series Reconstruction In regard to [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Warp block. 6 [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (2 more)
Figure 5
Figure 5. Figure 5: Visualization of the three classes of [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: The average distance from the MFD target data to their nearest neighbors in the source [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

41 extracted references · 34 canonical work pages

  1. [1]

    Time series domain adaptation via sparse associative structure alignment

    Ruichu Cai, Jiawei Chen, Zijian Li, Wei Chen, Keli Zhang, Junjian Ye, Zhuozhang Li, Xiaoyan Yang, and Zhenjie Zhang. Time series domain adaptation via sparse associative structure alignment. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 35, pages 6859--6867, 2021

    work page 2021

  2. [2]

    Homm: Higher-order moment matching for unsupervised domain adaptation

    Chao Chen, Zhihang Fu, Zhihong Chen, Sheng Jin, Zhaowei Cheng, Xinyu Jin, and Xian-Sheng Hua. Homm: Higher-order moment matching for unsupervised domain adaptation. In Proceedings of the AAAI conference on artificial intelligence, volume 34, pages 3422--3429, 2020

    work page 2020

  3. [3]

    Domain-adversarial training of neural networks

    Yaroslav Ganin, Evgeniya Ustinova, Hana Ajakan, Pascal Germain, Hugo Larochelle, Fran c ois Laviolette, Mario March, and Victor Lempitsky. Domain-adversarial training of neural networks. Journal of machine learning research, 17 0 (59): 0 1--35, 2016

    2016

  4. [4]

    Deep reconstruction-classification networks for unsupervised domain adaptation

    Muhammad Ghifary, W Bastiaan Kleijn, Mengjie Zhang, David Balduzzi, and Wen Li. Deep reconstruction-classification networks for unsupervised domain adaptation. In Computer Vision--ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11--14, 2016, Proceedings, Part IV 14, pages 597--613. Springer, 2016

    work page 2016

  5. [5]

    Sotta: Robust test-time adaptation on noisy data streams

    Taesik Gong, Yewon Kim, Taeckyung Lee, Sorn Chottananurak, and Sung-Ju Lee. Sotta: Robust test-time adaptation on noisy data streams. Advances in Neural Information Processing Systems, 36, 2024

    work page 2024

  6. [6]

    Sparse binary transformers for multivariate time series modeling

    Matt Gorbett, Hossein Shirazi, and Indrakshi Ray. Sparse binary transformers for multivariate time series modeling. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pages 544--556, 2023

    work page 2023

  7. [7]

    Domain adaptation for time series under feature and label shifts

    Huan He, Owen Queen, Teddy Koker, Consuelo Cuevas, Theodoros Tsiligkaridis, and Marinka Zitnik. Domain adaptation for time series under feature and label shifts. In International Conference on Machine Learning, pages 12746--12774. PMLR, 2023

    work page 2023

  8. [8]

    Swl-adapt: An unsupervised domain adaptation model with sample weight learning for cross-user wearable human activity recognition

    Rong Hu, Ling Chen, Shenghuan Miao, and Xing Tang. Swl-adapt: An unsupervised domain adaptation model with sample weight learning for cross-user wearable human activity recognition. In Proceedings of the AAAI Conference on artificial intelligence, volume 37, pages 6012--6020, 2023

    work page 2023

Show all 41 references
  1. [9]

    Sf-adapter: Computational-efficient source-free domain adaptation for human activity recognition

    Hua Kang, Qingyong Hu, and Qian Zhang. Sf-adapter: Computational-efficient source-free domain adaptation for human activity recognition. Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies, 7 0 (4): 0 1--23, 2024

    2024

  2. [10]

    Time series as images: Vision transformer for irregularly sampled time series

    Zekun Li, Shiyang Li, and Xifeng Yan. Time series as images: Vision transformer for irregularly sampled time series. Advances in Neural Information Processing Systems, 36, 2024

    2024

  3. [11]

    Do we really need to access the source data? source hypothesis transfer for unsupervised domain adaptation

    Jian Liang, Dapeng Hu, and Jiashi Feng. Do we really need to access the source data? source hypothesis transfer for unsupervised domain adaptation. In International conference on machine learning, pages 6028--6039. PMLR, 2020

    2020

  4. [12]

    Ttn: A domain-shift aware batch normalization in test-time adaptation

    Hyesu Lim, Byeonggeun Kim, Jaegul Choo, and Sungha Choi. Ttn: A domain-shift aware batch normalization in test-time adaptation. arXiv preprint arXiv:2302.05155, 2023

    2023 arXiv

  5. [13]

    Timesurl: Self-supervised contrastive learning for universal time series representation learning

    Jiexi Liu and Songcan Chen. Timesurl: Self-supervised contrastive learning for universal time series representation learning. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 13918--13926, 2024

    2024

  6. [14]

    Adversarial spectral kernel matching for unsupervised time series domain adaptation

    Qiao Liu and Hui Xue. Adversarial spectral kernel matching for unsupervised time series domain adaptation. In IJCAI, pages 2744--2750, 2021

    2021

  7. [15]

    Conditional adversarial domain adaptation

    Mingsheng Long, Zhangjie Cao, Jianmin Wang, and Michael I Jordan. Conditional adversarial domain adaptation. Advances in neural information processing systems, 31, 2018

    2018

  8. [16]

    Fft-trans: Enhancing robustness in mechanical fault diagnosis with fourier transform-based transformer under noisy conditions

    Xiaoyu Luo, Huan Wang, Te Han, and Ying Zhang. Fft-trans: Enhancing robustness in mechanical fault diagnosis with fourier transform-based transformer under noisy conditions. IEEE Transactions on Instrumentation and Measurement, 2024

    2024

  9. [17]

    Mhccl: masked hierarchical cluster-wise contrastive learning for multivariate time series

    Qianwen Meng, Hangwei Qian, Yong Liu, Lizhen Cui, Yonghui Xu, and Zhiqi Shen. Mhccl: masked hierarchical cluster-wise contrastive learning for multivariate time series. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 37, pages 9153--9161, 2023

    2023

  10. [18]

    Formertime: hierarchical multi-scale representations for multivariate time series classification

    Cheng Mingyue, Liu Qi, Liu Zhiding, Li Zhi, Luo Yucong, and Chen Enhong. Formertime: hierarchical multi-scale representations for multivariate time series classification. arXiv e-prints, 2023

    2023

  11. [19]

    Towards stable test-time adaptation in dynamic wild world

    Shuaicheng Niu, Jiaxiang Wu, Yifan Zhang, Zhiquan Wen, Yaofo Chen, Peilin Zhao, and Mingkui Tan. Towards stable test-time adaptation in dynamic wild world. arXiv preprint arXiv:2302.12400, 2023

    2023 arXiv

  12. [20]

    Towards interpretable sleep stage classification using cross-modal transformers

    Jathurshan Pradeepkumar, Mithunjha Anandakumar, Vinith Kugathasan, Dhinesh Suntharalingham, Simon L Kappel, Anjula C De Silva, and Chamira US Edussooriya. Towards interpretable sleep stage classification using cross-modal transformers. IEEE Transactions on Neural Systems and R...

    2024

  13. [21]

    Adaptive intermediate class-wise distribution alignment: a universal domain adaptation and generalization method for machine fault diagnosis

    Quan Qian, Jun Luo, and Yi Qin. Adaptive intermediate class-wise distribution alignment: a universal domain adaptation and generalization method for machine fault diagnosis. IEEE transactions on neural networks and learning systems, 2024

    2024

  14. [22]

    Adatime: A benchmarking suite for domain adaptation on time series data

    Mohamed Ragab, Emadeldeen Eldele, Wee Ling Tan, Chuan-Sheng Foo, Zhenghua Chen, Min Wu, Chee-Keong Kwoh, and Xiaoli Li. Adatime: A benchmarking suite for domain adaptation on time series data. ACM Transactions on Knowledge Discovery from Data, 17 0 (8): 0 1--18, 2023 a

    2023

  15. [23]

    Source-free domain adaptation with temporal imputation for time series data

    Mohamed Ragab, Emadeldeen Eldele, Min Wu, Chuan-Sheng Foo, Xiaoli Li, and Zhenghua Chen. Source-free domain adaptation with temporal imputation for time series data. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pages 1989--1998, 2023 b

    1989

  16. [24]

    On minimum discrepancy estimation for deep domain adaptation

    Mohammad Mahfujur Rahman, Clinton Fookes, Mahsa Baktashmotlagh, and Sridha Sridharan. On minimum discrepancy estimation for deep domain adaptation. Domain Adaptation for Visual Understanding, pages 81--94, 2020

    2020

  17. [25]

    Correlation alignment for unsupervised domain adaptation

    Baochen Sun, Jiashi Feng, and Kate Saenko. Correlation alignment for unsupervised domain adaptation. Domain adaptation in computer vision applications, pages 153--171, 2017

    2017

  18. [26]

    A universal multi-source domain adaptation method with unsupervised clustering for mechanical fault diagnosis under incomplete data

    Jinghui Tian, Dongying Han, Hamid Reza Karimi, Yu Zhang, and Peiming Shi. A universal multi-source domain adaptation method with unsupervised clustering for mechanical fault diagnosis under incomplete data. Neural Networks, 173: 0 106167, 2024

    2024

  19. [27]

    Deep domain confusion: Maximizing for domain invariance

    Eric Tzeng, Judy Hoffman, Ning Zhang, Kate Saenko, and Trevor Darrell. Deep domain confusion: Maximizing for domain invariance. arXiv preprint arXiv:1412.3474, 2014

    2014 arXiv

  20. [28]

    Neural discrete representation learning

    Aaron Van Den Oord, Oriol Vinyals, et al. Neural discrete representation learning. Advances in neural information processing systems, 30, 2017

    2017

  21. [29]

    Tent: Fully test-time adaptation by entropy minimization

    Dequan Wang, Evan Shelhamer, Shaoteng Liu, Bruno Olshausen, and Trevor Darrell. Tent: Fully test-time adaptation by entropy minimization. arXiv preprint arXiv:2006.10726, 2020

    2006 arXiv

  22. [30]

    When: A wavelet-dtw hybrid attention network for heterogeneous time series analysis

    Jingyuan Wang, Chen Yang, Xiaohan Jiang, and Junjie Wu. When: A wavelet-dtw hybrid attention network for heterogeneous time series analysis. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pages 2361--2373, 2023

    2023

  23. [31]

    Pond: Multi-source time series domain adaptation with information-aware prompt tuning

    Junxiang Wang, Guangji Bai, Wei Cheng, Zhengzhang Chen, Liang Zhao, and Haifeng Chen. Pond: Multi-source time series domain adaptation with information-aware prompt tuning. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pages 3140--3151, 2024

    2024

  24. [32]

    Multi-source deep domain adaptation with weak supervision for time-series sensor data

    Garrett Wilson, Janardhan Rao Doppa, and Diane J Cook. Multi-source deep domain adaptation with weak supervision for time-series sensor data. In Proceedings of the 26th ACM SIGKDD international conference on knowledge discovery & data mining, pages 1768--1778, 2020

    2020

  25. [33]

    Calda: Improving multi-source time series domain adaptation with contrastive adversarial learning

    Garrett Wilson, Janardhan Rao Doppa, and Diane J Cook. Calda: Improving multi-source time series domain adaptation with contrastive adversarial learning. IEEE transactions on pattern analysis and machine intelligence, 2023

    2023

  26. [34]

    Privacy-preserving domain adaptation for motor imagery-based brain-computer interfaces

    Kun Xia, Lingfei Deng, Wlodzislaw Duch, and Dongrui Wu. Privacy-preserving domain adaptation for motor imagery-based brain-computer interfaces. IEEE Transactions on Biomedical Engineering, 69 0 (11): 0 3365--3376, 2022

    2022

  27. [35]

    Channel attention for sensor-based activity recognition: embedding features into all frequencies in dct domain

    Shige Xu, Lei Zhang, Yin Tang, Chaolei Han, Hao Wu, and Aiguo Song. Channel attention for sensor-based activity recognition: embedding features into all frequencies in dct domain. IEEE Transactions on Knowledge and Data Engineering, 35 0 (12): 0 12497--12512, 2023

    2023

  28. [36]

    Exploiting the intrinsic neighborhood structure for source-free domain adaptation

    Shiqi Yang, Joost Van de Weijer, Luis Herranz, Shangling Jui, et al. Exploiting the intrinsic neighborhood structure for source-free domain adaptation. Advances in neural information processing systems, 34: 0 29393--29405, 2021

    2021

  29. [37]

    Attracting and dispersing: A simple approach for source-free domain adaptation

    Shiqi Yang, Shangling Jui, Joost van de Weijer, et al. Attracting and dispersing: A simple approach for source-free domain adaptation. Advances in Neural Information Processing Systems, 35: 0 5802--5815, 2022

    2022

  30. [38]

    Deep generative domain adaptation with temporal relation attention mechanism for cross-user activity recognition

    Xiaozhou Ye, I Kevin, and Kai Wang. Deep generative domain adaptation with temporal relation attention mechanism for cross-user activity recognition. Pattern Recognition, 156: 0 110811, 2024

    2024

  31. [39]

    Self-supervised contrastive pre-training for time series via time-frequency consistency

    Xiang Zhang, Ziyuan Zhao, Theodoros Tsiligkaridis, and Marinka Zitnik. Self-supervised contrastive pre-training for time series via time-frequency consistency. Advances in Neural Information Processing Systems, 35: 0 3988--4003, 2022

    2022

  32. [40]

    Multi-modal sleep stage classification with two-stream encoder-decoder

    Zhao Zhang, Bor-Shyh Lin, Chih-Wei Peng, and Bor-Shing Lin. Multi-modal sleep stage classification with two-stream encoder-decoder. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 32: 0 2096--2105, 2024

    2024

  33. [41]

    Deep representation-based domain adaptation for nonstationary eeg classification

    He Zhao, Qingqing Zheng, Kai Ma, Huiqi Li, and Yefeng Zheng. Deep representation-based domain adaptation for nonstationary eeg classification. IEEE Transactions on Neural Networks and Learning Systems, 32 0 (2): 0 535--545, 2020

    2020

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.