Pith. sign in

REVIEW 4 major objections 5 minor 40 references

TransPL: VQ-Code Transition Matrices for Pseudo-Labeling of Time Series Unsupervised Domain Adaptation

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read TransPL pseudo-labels time-series target data using class- and channel-wise transition matrices over vector-quantized codes, and reports the best adaptation accuracy on all four UDA benchmarks.

desk verdict TransPL is a genuinely new pseudo-labeling mechanism for time-series UDA with solid average gains, but the transferability of its class-conditional transition matrices is asserted rather than shown, and the statistics need work. read the letter →

arxiv 2505.09955 v1 pith:GV5IOW2R submitted 2025-05-15 cs.LG cs.AI

Jaeho Kim , Seulki Lee This is my paper · ORCID
classification cs.LGcs.AI
keywords unsuperviseddomainadaptationtimeseriesclassificationpseudo-labelingvectorquantizationtransitionmatrixchannelalignmentoptimaltransportweaksupervision
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

TransPL is an unsupervised domain adaptation method for time series that generates pseudo-labels from explicit temporal transition statistics rather than from a black-box classifier. The paper's central claim is that modeling the source joint distribution $P(\mathbf{X}, y)$ through class- and channel-wise transition matrices over vector-quantized (VQ) coarse codes—counting how often one learned code follows another within each class and each sensor—is enough to pseudo-label the unlabeled target domain via Bayes' rule. Each channel votes for a class through its class-conditional transition likelihood, and the votes are weighted by how well that channel's source and target transition matrices align, measured by optimal transport. On four time-series benchmarks (UCIHAR, WISDM, HHAR, PTB) the paper reports the best adaptation accuracy and macro-F1 among all compared methods, with average gains of 11.4% accuracy and 12.2% MF1 over no adaptation and 3.0% accuracy and 2.9% MF1 over SHOT. Weak supervision, in the form of a known label prior, plugs directly into the Bayes formulation and improves results further.

What carries the argument

The load-bearing object is the coarse-code transition matrix: an empirical Markov transition matrix $P_{cl}^S \in \mathbb{R}^{K \times D \times n_c \times n_c}$ built by counting transitions between vector-quantized coarse codes of time-series patches, per class and per channel. The supporting machinery is a two-level residual VQ structure (coarse codebook with $n_c=8$ codes, fine codebook with $n_f=64$ codes) that keeps transition counting tractable while preserving reconstruction; an optimal-transport channel alignment score $w_d$ that down-weights sensors whose source-to-target transition rows have drifted, using cosine distance between codes as the transport cost; and Bayes' rule with a label prior $p(k)$ that computes channel-wise class posteriors from the class-conditional transition likelihoods. Together these turn pseudo-labeling into a transparent likelihood computation: the transition matrices are the model of $P(\mathbf{X}, y)$, the transport cost is the measure of domain shift, and the posterior is the pseudo-label.

What would settle it

Construct or select a source-target pair in which one class's temporal dynamics change (for example, an activity performed with a different rhythm or a sensor whose mounting changes) while another class keeps its pattern, and compare TransPL's per-class pseudo-label accuracy against a target-clustering baseline such as SHOT. If the shifted class's pseudo-label accuracy drops below the clustering baseline even though the channel alignment scores stay high, the central claim that source transition matrices transfer across domains is contradicted.

Watch

Extended reading notes

Core claim

On the paper's own terms, TransPL establishes that a generative, explainable pseudo-labeling rule can outperform discriminative pseudo-labeling in time-series UDA. After training a patch encoder, decoder, coarse codebook ($n_c=8$) and fine codebook ($n_f=64$) on labeled source data, it freezes the model and infers coarse codes for both domains. From the labeled source it builds a class-wise transition matrix $P_{cl}^S \in \mathbb{R}^{K \times D \times n_c \times n_c}$ whose entries are empirical one-step Markov transition counts between coarse codes, per class and per channel; from both domains it builds channel-wise transition matrices $P_{ch}^S$ and $P_{ch}^T$. For a target sequence it computes the log-likelihood of its coarse code sequence under each class's transition matrix (Equation 7), converts this to a channel-wise class posterior with Bayes' rule (Equation 3), and averages the posteriors across channels with weights $w_d$ derived from the earth mover's distance between source and target transition rows, with code similarity encoded in the transport cost. The paper reports this pseudo-labeling rule as the best across the four datasets, with pseudo-label accuracy gains of 6.1% (accuracy) and 4.9% (MF1) over the best baselines, and shows that the learned transition matrices are interpretable: they reveal which classes have similar temporal dynamics and which channels have shifted between domains.

Load-bearing premise

The load-bearing premise is that the per-class temporal transition patterns counted on the source domain remain valid for target sequences of the same class, so the source class-wise transition matrices can serve as target class-conditional likelihoods; if a class's transition signature drifts between domains while its channel marginals look similar, the pseudo-labels will be miscalibrated.

Editorial extensions

If this is right

  • If the reported results hold, time-series UDA pseudo-labeling no longer needs a source classifier's softmax scores or target clustering; a countable transition model over VQ codes suffices, making the pseudo-labeling process inspectable.
  • Weak supervision in the form of a known target label distribution can be injected as the prior $p(k)$ in Equation (3), turning a heuristic KL-divergence penalty into a principled Bayes update; the paper reports gains in all four datasets from this prior.
  • Because pseudo-labels are transition-based rather than amplitude-based, the method should stay accurate under amplitude-level sensor shifts as long as each class's coarse-code transition signature persists; the paper demonstrates this invariance on the ECG and HAR benchmarks.
  • Channel alignment weights computed from transition-matrix transport distances flag which sensors are drifting, giving a per-dataset explanation of what needs adaptation.
  • The coarse/fine codebook design ($n_c=8$, $n_f=64$) yields zero dead codes and tractable $8 \times 8$ transition matrices, so the approach scales to long sequences without expensive density estimation.

Reading between the lines

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

  • An unstated consequence is that Equation (7) uses the source class-wise transition matrix as the likelihood for target sequences, so per-class drift in temporal transition patterns that is not a whole-channel shift will mis-calibrate pseudo-labels; the channel weights $w_d$ cannot repair this because they down-weight entire channels, not specific classes within a channel. A natural extension is es
  • The transport cost between source and target channel transition matrices is itself a direct measure of how much each sensor's dynamics changed, which could serve as a dataset-level UDA difficulty score or a sensor-selection criterion.
  • The same transition likelihoods could be used for anomaly or out-of-distribution detection on time series: a sequence whose code transitions are unlikely under all class transition matrices would be flagged as novel.
  • Because the pseudo-label is a channel-weighted posterior, the top-$r$ confidence selection in Section 5.2 could be replaced by per-sample posterior entropy thresholds, which may improve robustness for classes with sparse transition evidence.
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 / 5 minor

Summary. The paper proposes TransPL, a pseudo-labeling method for time-series unsupervised domain adaptation (UDA). A VQ-VAE is trained on labeled source data with coarse and fine codebooks; from the coarse codes the method builds class-wise transition matrices per channel and channel-wise transition matrices for source and target. Target sequences are pseudo-labeled by Bayes' rule using class-conditional likelihoods from Eq. (7), a label prior, and optimal-transport-based channel alignment weights w_d from Eq. (6). The adapted model is fine-tuned with the pseudo-labeled target set. Experiments on UCIHAR, WISDM, HHAR, and PTB report mean accuracy and macro-F1 improvements over baselines, plus ablations and weak-supervision experiments.

Significance. If the results hold, TransPL would be a useful and interpretable new mechanism for time-series UDA, distinguished by explicit temporal-transition modeling via VQ codes and by a principled channel weighting scheme. The paper provides a reproducible code repository, includes full per-pair tables, ablations of channel alignment and weak supervision, and several focused analyses (permutation entropy of coarse vs. fine codes, channel corruption experiments) that isolate the contribution of its modules. However, the central empirical claims are weakened by the absence of variance or statistical tests, by hyperparameter selection that uses target labels, and by per-pair failures that contradict the word 'consistent' in the claimed advantage; these need to be resolved before the claimed state-of-the-art status is established.

major comments (4)
  1. [Section 4.2 and Eq. (7)] The method's core assumption is that source class-wise transition matrices serve as target class-conditional likelihoods. The justification given in Section 4.2 follows from the covariate-shift assumption P(y|X)^S = P(y|X)^T, but that assumption only fixes the class posterior; it does not imply that P(X^d|y=k) or the coarse-code transition probabilities within class k are invariant across domains. Since Eq. (7) computes target likelihoods directly from P^S_{cl,k}, any per-class transition drift systematically mis-calibrates the Bayes posterior, and the channel weights of Eq. (6) only rescale whole channels. The paper should validate this transfer assumption, for example by comparing class-wise transition matrices estimated from labeled target test data against the source matrices and correlating per-pair distances with per-pair pseudo-label errors in Table 3, or by reporting whether the large per-pair SHOT losses in Tables 9–16 correspond to large transition drift.
  2. [Section 6 and Table 1] The codebook size selection in Table 1 uses Pseudo-Label Accuracy on the target training set, which requires target labels. This is a UDA violation: the numbers of coarse and fine codes n_c and n_f, set to 8 and 64 in Section 6, are chosen with access to target ground truth, and the baseline comparisons are therefore not obtained under the same information constraint. Please re-select hyperparameters using only source validation data (as stated in the evaluation protocol) or an unsupervised criterion such as reconstruction error, and report whether the selected n_c, n_f and the other hyperparameters of Table 17 remain unchanged.
  3. [Section 7.1 and Tables 9–16] The phrase 'consistent advantage over SHOT' is not supported at the per-pair level. For example, TransPL's accuracy is 66.4 vs SHOT 90.3 on UCIHAR 12→16, 48.5 vs 78.8 on WISDM 6→19, 54.5 vs 83.3 on WISDM 28→4, and 75.4 vs 86.5 on HHAR 4→5; the averages in Table 2 hide these large degradations. With only 10 pairs per dataset and no variance or significance test, the claim that TransPL 'demonstrates the best adaptation performance across all four datasets' is stronger than the evidence. Please report per-pair wins/losses, paired tests (e.g., Wilcoxon signed-rank), and standard deviations or confidence intervals.
  4. [Eq. (7)] The class-conditional likelihood computation does not specify how zero-count transitions in P^S_{cl,k} are handled. For sequences of length N≈16–20 and a codebook of 8 states, many transitions will be unobserved in the source class-wise counts, making the log-likelihood in Eq. (7) −∞ unless a smoothing or Laplace correction is applied. The paper only mentions epsilon smoothing for the channel-wise TMs in Section 5.1; please state the exact smoothing used for the class-wise TMs, since it directly affects the pseudo-label confidences and the r_top selection in Section 5.2.
minor comments (5)
  1. [Abstract] The '6.1% accuracy improvement, 4.9% F1 improvement' is the gain in pseudo-label accuracy from Table 3, not the final adaptation performance gain from Table 2; please state this explicitly to avoid misreading.
  2. [Table 2] The RAINCOAT row appears garbled ('RAINCOAT58.6 61.1 47.3 63.0 58.6 61.1 59.8 41.2'); the values seem inconsistent with the individual results in Appendix D and with the averages shown elsewhere.
  3. [Eq. (3)] Since the channel weights w_d are not renormalized to sum to one, a channel with very low alignment reduces the overall pseudo-label magnitude; please specify whether this is intentional or whether the weights should be normalized.
  4. [Section 5.1] The optimal transport cost matrix M in Eq. (5) is defined on coarse codes from the shared source codebook; please state explicitly that the same codebook is used for both domains and that source and target codes are aligned in index space.
  5. [Appendix E and Table 17] Please report how σ, τ, r_top, and the source/adaptation learning rates were selected (source validation or target labels) to allow reproducibility.

Circularity Check

1 steps flagged · score 2.0 of 10

The core Bayes/transition-matrix derivation is self-contained; the only circularity-adjacent issue is a target-label-dependent model-selection leak in Table 1, which is a correctness concern rather than a definitional circularity.

  1. fitted input called prediction [Section 4.1 (Table 1) and Section 6 (Implementation); cf. Section 7.2]
    "Table 1: Comparison of single VQV AE (top three rows) versus our residual coarse-fine VQ codebook configurations (bottom rows) on HHAR task. Results are averages of 10 source-target pairs. ... We selected the number of coarse and fine codes to nc = 8, nf = 64 for all tasks. ... We report the accuracy of the constructed pseudo label of TransPL in Table 3 over the unlabeled target domain's training set ... In practice, the pseudo label's performance is not used or known during model training as we do not have access to the true labels of the target domain."

    The codebook size is selected by maximizing pseudo-label accuracy (PL Acc) on the HHAR target training set, and the same selected configuration (nc=8, nf=64) is then used to produce all reported adaptation and pseudo-label numbers. Since pseudo-label accuracy itself is measured against true target training labels, this is model selection on target-label information: the reported strong pseudo-label and adaptation results are partly forced by choosing hyperparameters that already score well on the target domain. This is a target-label leakage in the experimental pipeline, not a mathematical equivalence inside Eq. (7); the derivation itself does not define the prediction as the fitted quantity.

full rationale

The central claim of TransPL is not circular by construction. Equation (2) estimates transition probabilities from labeled source counts; Equation (7) computes class-conditional likelihoods by applying those frozen source TMs to target coarse-code sequences; Equation (3) forms pseudo-labels via Bayes' rule with channel alignment weights. None of these steps defines the target pseudo-label as the fitted input, and evaluation is performed on held-out target test data using the source risk protocol. The covariate-shift justification in Section 4.2 is an empirical assumption about class-conditional transfer, not a reduction of the output to the input; the paper's own Section 8 limitation honestly flags the channel-weighting failure mode. There is no load-bearing self-citation chain: the only same-author citation (Kim et al. 2024) appears in a limitation remark, and the encoder reference (Gui et al. 2024) is an external architectural choice. The one genuine flaw is the Table 1 codebook selection based on pseudo-label accuracy on the target training set, which uses target labels and leaks into the final reported configuration. That is a model-selection leak and a correctness issue, but it does not make the derivation itself definitionally circular; hence the low score of 2.

Assumptions & free parameters 7 free parameters · 6 assumptions · 0 invented entities

The central method depends on a small set of tuned hyperparameters and on the assumption that class-conditional transition dynamics survive domain shift. No new physical or mathematical entities are postulated; the coarse and fine codebooks are learned representations rather than invented entities.

free parameters (7)
  • coarse codebook size n_c = 8
    Chosen based on Table 1, which reports pseudo-label accuracy on the target training set for different codebook sizes; used in all tasks.
  • fine codebook size n_f = 64
    Selected together with n_c=8 in Table 1 to maximize pseudo-label accuracy with zero dead codes.
  • patch length m = 8 (UCIHAR, WISDM, HHAR), 15 (PTB)
    Selected per dataset; Appendix F reports performance across patch lengths, implying selection based on validation or target performance.
  • RBF bandwidth sigma = 0.2 (UCIHAR), 0.1 (WISDM), 0.1 (HHAR), 0.2 (PTB)
    Per-task hyperparameter in Eq. 6 that controls channel alignment weighting; tuned per dataset (Table 17).
  • prior temperature tau = 1.0, 2.0, 5.0, 1.0
    Modulates weak-supervision prior strength in the log-prior term; tuned per dataset (Table 17).
  • top selection ratio r_top = 0.5, 0.2, 0.2, 0.7
    Fraction of highest-confidence pseudo-labels used for fine-tuning; tuned per dataset (Table 17).
  • commitment beta = 0.25
    Standard VQ-VAE commitment cost, set by hand.
assumptions (6)
  • domain assumption Covariate shift: P(y|X)^S = P(y|X)^T with P_X^S != P_X^T.
    Stated in Section 3 and used to justify replacing target class-conditional likelihoods with source transition matrices.
  • domain assumption Within a class, coarse-code transition distributions transfer from source to target.
    Section 4.2 asserts that under covariate shift 'temporal sequences from the same class should have similar coarse code transition patterns'; this is stronger than covariate shift because p(X|y) may shift.
  • standard math Coarse code sequences form a first-order Markov chain.
    Used in Eq. 2 and Eq. 7; a modeling approximation for temporal dependencies.
  • domain assumption Coarse codes capture global temporal patterns sufficient for classification while fine codes capture residual detail.
    Supported empirically by permutation entropy in Figure 2, but the paper states in Section 8 this hierarchy is not mathematically regularized.
  • domain assumption The VQ-VAE encoder and codebooks trained on source reconstruct target patches meaningfully.
    Pseudo-labeling in Section 5.1 infers target coarse codes using the source-trained encoder and codebooks without adaptation.
  • domain assumption Pseudo-labels from the top-r_top confidence subset are reliable enough to fine-tune the full model.
    Section 5.2 selects a reliable subset; this is the standard pseudo-labeling premise and is not independently verified beyond the empirical gains.

how reviews work

0 comments
Cite this review

Pith. "Pith review of TransPL: VQ-Code Transition Matrices for Pseudo-Labeling of Time Series Unsupervised Domain Adaptation." pith.science (2026). https://pith.science/paper/GV5IOW2R

@misc{pith2026250509955,
  author       = {Pith},
  title        = {Pith review of: TransPL: VQ-Code Transition Matrices for Pseudo-Labeling of Time Series Unsupervised Domain Adaptation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GV5IOW2R}},
  note         = {Machine review of arXiv:2505.09955}
}
abstract

Unsupervised domain adaptation (UDA) for time series data remains a critical challenge in deep learning, with traditional pseudo-labeling strategies failing to capture temporal patterns and channel-wise shifts between domains, producing sub-optimal pseudo-labels. As such, we introduce TransPL, a novel approach that addresses these limitations by modeling the joint distribution $P(\mathbf{X}, y)$ of the source domain through code transition matrices, where the codes are derived from vector quantization (VQ) of time series patches. Our method constructs class- and channel-wise code transition matrices from the source domain and employs Bayes' rule for target domain adaptation, generating pseudo-labels based on channel-wise weighted class-conditional likelihoods. TransPL offers three key advantages: explicit modeling of temporal transitions and channel-wise shifts between different domains, versatility towards different UDA scenarios (e.g., weakly-supervised UDA), and explainable pseudo-label generation. We validate TransPL's effectiveness through extensive analysis on four time series UDA benchmarks and confirm that it consistently outperforms state-of-the-art pseudo-labeling methods by a strong margin (6.1% accuracy improvement, 4.9% F1 improvement), while providing interpretable insights into the domain adaptation process through its learned code transition matrices.

Figures

Figures reproduced from arXiv: 2505.09955 by the authors.

Figure 1
Figure 1. Overall scheme of TransPL. (A) Source Training: We first train the whole model architecture, i.e., encoder, decoder, two VQ codebooks (coarse and fine codebook), and classifier, using the labeled source domain data. A [CLS] token is appended to the input patches and is used as the input to the classifier. (B) Coarse Code Transition Matrices (TM): The trained encoder and coarse codebook infer coarse codes from both t… view at source ↗
Figure 2
Figure 2. A. Average PE: The average permutation entropy (PE) score for the reconstructed coarse and fine VQ codes. PE is a measure of temporal complexity (Bandt & Pompe, 2002), where a lower PE indicates a simpler pattern. Across all time series benchmarks (i.e. UCIHAR, HHAR, WISDM), the coarse code exhibits lower PE compared to the fine code. B. Visualization: The reconstructed coarse and fine VQ codes shows that the coarse… view at source ↗
Figure 3
Figure 3. Class conditional likelihood visualization in UCI. [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Channel alignment (CA) analysis. We visualized the latent representation of each channels’s [CLS] token and observe that the degree of shift differs between channels. Specifically, channel 3 shows higher degrees of shift between source and target compared to channel 1 …
Figure 5
Figure 5. Figure 5: Visualization of the full coarse and fine codes in HHAR task. [PITH_FULL_IMAGE:figures/full_fig_p013_5.png]
Figure 6
Figure 6. Figure 6: Visualization of the ranks of w6 for the UCIHAR task. Our results demonstrate that as noise magnitude increases in the sixth channel, its rank consistently decreases across all tested pairs, confirming that our channel alignment module successfully identifies channels …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 22 canonical work pages

  1. [1]

    A., and Dyer, E

    Ahad, N., Davenport, M. A., and Dyer, E. L. Time series domain adaptation via channel-selective representation alignment. Transactions on Machine Learning Research, 2025

    work page 2025

  2. [2]

    L., et al

    Anguita, D., Ghio, A., Oneto, L., Parra, X., Reyes-Ortiz, J. L., et al. A public domain dataset for human activity recognition using smartphones. In Esann, volume 3, pp.\ 3, 2013

    work page 2013

  3. [3]

    and Pompe, B

    Bandt, C. and Pompe, B. Permutation entropy: a natural complexity measure for time series. Physical review letters, 88 0 (17): 0 174102, 2002

    work page 2002

  4. [4]

    Bishop, C. M. and Nasrabadi, N. M. Pattern recognition and machine learning, volume 4. Springer, 2006

    2006

  5. [5]

    Nutzung der ekg-signaldatenbank cardiodat der ptb \"u ber das internet

    Bousseljot, R., Kreiseler, D., and Schnabel, A. Nutzung der ekg-signaldatenbank cardiodat der ptb \"u ber das internet. 1995

    work page 1995

  6. [6]

    Brockwell, P. J. and Davis, R. A. Introduction to time series and forecasting. Springer, 2002

    work page 2002

  7. [7]

    Time series domain adaptation via sparse associative structure alignment

    Cai, R., Chen, J., Li, Z., Chen, W., Zhang, K., Ye, J., Li, Z., Yang, X., and Zhang, Z. Time series domain adaptation via sparse associative structure alignment. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 35, pp.\ 6859--6867, 2021

    work page 2021

  8. [8]

    Learning phrase representations using rnn encoder-decoder for statistical machine translation

    Cho, K., Van Merri \"e nboer, B., Gulcehre, C., Bahdanau, D., Bougares, F., Schwenk, H., and Bengio, Y. Learning phrase representations using rnn encoder-decoder for statistical machine translation. arXiv preprint arXiv:1406.1078, 2014

    arXiv 2014

Show all 40 references
  1. [9]

    Z., Boisbunon, A., Chambon, S., Chapel, L., Corenflos, A., Fatras, K., Fournier, N., et al

    Flamary, R., Courty, N., Gramfort, A., Alaya, M. Z., Boisbunon, A., Chambon, S., Chapel, L., Corenflos, A., Fatras, K., Fournier, N., et al. Pot: Python optimal transport. Journal of Machine Learning Research, 22 0 (78): 0 1--8, 2021

    2021

  2. [10]

    Domain-adversarial training of neural networks

    Ganin, Y., Ustinova, E., Ajakan, H., Germain, P., Larochelle, H., Laviolette, F., March, M., and Lempitsky, V. Domain-adversarial training of neural networks. Journal of machine learning research, 17 0 (59): 0 1--35, 2016

    2016

  3. [11]

    Vector quantization pretraining for eeg time series with random projection and phase alignment

    Gui, H., Li, X., and Chen, X. Vector quantization pretraining for eeg time series with random projection and phase alignment. In International Conference on Machine Learning, pp.\ 16731--16750. PMLR, 2024

    2024

  4. [12]

    Domain adaptation for time series under feature and label shifts

    He, H., Queen, O., Koker, T., Cuevas, C., Tsiligkaridis, T., and Zitnik, M. Domain adaptation for time series under feature and label shifts. In International Conference on Machine Learning, pp.\ 12746--12774. PMLR, 2023 a

    2023

  5. [13]

    He, Q.-Q., Siu, S. W. I., and Si, Y.-W. Attentive recurrent adversarial domain adaptation with top-k pseudo-labeling for time series classification. Applied Intelligence, 53 0 (11): 0 13110--13129, 2023 b

    2023

  6. [14]

    Long short-term memory

    Hochreiter, S. Long short-term memory. Neural Computation MIT-Press, 1997

    1997

  7. [15]

    Multi-task learning using uncertainty to weigh losses for scene geometry and semantics

    Kendall, A., Gal, Y., and Cipolla, R. Multi-task learning using uncertainty to weigh losses for scene geometry and semantics. In Proceedings of the IEEE conference on computer vision and pattern recognition, pp.\ 7482--7491, 2018

    2018

  8. [16]

    Cafo: Feature-centric explanation on time series classification

    Kim, J., Hahn, S.-J., Hwang, Y., Lee, J., and Lee, S. Cafo: Feature-centric explanation on time series classification. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pp.\ 1372--1382, 2024

    2024

  9. [17]

    R., Weiss, G

    Kwapisz, J. R., Weiss, G. M., and Moore, S. A. Activity recognition using cell phone accelerometers. ACM SigKDD Explorations Newsletter, 12 0 (2): 0 74--82, 2011

    2011

  10. [18]

    Vector quantized time series generation with a bidirectional prior model

    Lee, D., Malacarne, S., and Aune, E. Vector quantized time series generation with a bidirectional prior model. arXiv preprint arXiv:2303.04743, 2023

    2023 arXiv

  11. [19]

    Lee, D.-H. et al. Pseudo-label: The simple and efficient semi-supervised learning method for deep neural networks. In Workshop on challenges in representation learning, ICML, volume 3, pp.\ 896. Atlanta, 2013

    2013

  12. [20]

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

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

    2020

  13. [21]

    A symbolic representation of time series, with implications for streaming algorithms

    Lin, J., Keogh, E., Lonardi, S., and Chiu, B. A symbolic representation of time series, with implications for streaming algorithms. In Proceedings of the 8th ACM SIGMOD workshop on Research issues in data mining and knowledge discovery, pp.\ 2--11, 2003

    2003

  14. [22]

    Deja vu: Continual model generalization for unseen domains

    Liu, C., Wang, L., Lyu, L., Sun, C., Wang, X., and Zhu, Q. Deja vu: Continual model generalization for unseen domains. arXiv preprint arXiv:2301.10418, 2023 a

    2023 arXiv

  15. [23]

    Deja vu: Continual model generalization for unseen domains

    Liu, C., Wang, L., Lyu, L., Sun, C., Wang, X., and Zhu, Q. Deja vu: Continual model generalization for unseen domains. In The Eleventh International Conference on Learning Representations, 2023 b

    2023

  16. [24]

    Estimating intensity of physical activity: a comparison of wearable accelerometer and gyro sensors and 3 sensor locations

    Parkka, J., Ermes, M., Antila, K., Van Gils, M., Manttari, A., and Nieminen, H. Estimating intensity of physical activity: a comparison of wearable accelerometer and gyro sensors and 3 sensor locations. In 2007 29th annual international conference of the IEEE engineering in me...

    2007

  17. [25]

    and Werman, M

    Pele, O. and Werman, M. Fast and robust earth mover's distances. In 2009 IEEE 12th international conference on computer vision, pp.\ 460--467. IEEE, 2009

    2009

  18. [26]

    Computational optimal transport: With applications to data science

    Peyr \'e , G., Cuturi, M., et al. Computational optimal transport: With applications to data science. Foundations and Trends in Machine Learning , 11 0 (5-6): 0 355--607, 2019

    2019

  19. [27]

    L., Foo, C.-S., Chen, Z., Wu, M., Kwoh, C.-K., and Li, X

    Ragab, M., Eldele, E., Tan, W. L., Foo, C.-S., Chen, Z., Wu, M., Kwoh, C.-K., and Li, X. 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

    2023

  20. [28]

    M., Fookes, C., Baktashmotlagh, M., and Sridharan, S

    Rahman, M. M., Fookes, C., Baktashmotlagh, M., and Sridharan, S. On minimum discrepancy estimation for deep domain adaptation. Domain Adaptation for Visual Understanding, pp.\ 81--94, 2020

    2020

  21. [29]

    Asymmetric tri-training for unsupervised domain adaptation

    Saito, K., Ushiku, Y., and Harada, T. Asymmetric tri-training for unsupervised domain adaptation. In International conference on machine learning, pp.\ 2988--2997. PMLR, 2017

    2017

  22. [30]

    S., Kj rgaard, M

    Stisen, A., Blunck, H., Bhattacharya, S., Prentow, T. S., Kj rgaard, M. B., Dey, A., Sonne, T., and Jensen, M. M. Smart devices are different: Assessing and mitigatingmobile sensing heterogeneities for activity recognition. In Proceedings of the 13th ACM conference on embedded...

    2015

  23. [31]

    and Saenko, K

    Sun, B. and Saenko, K. Deep coral: Correlation alignment for deep domain adaptation. In Computer Vision--ECCV 2016 Workshops: Amsterdam, The Netherlands, October 8-10 and 15-16, 2016, Proceedings, Part III 14, pp.\ 443--450. Springer, 2016

    2016

  24. [32]

    Sun, S. et al. Caudits: Causal disentangled domain adaptation of multivariate time series. In Forty-first International Conference on Machine Learning, 2024

    2024

  25. [33]

    Totem: Tokenized time series embeddings for general time series analysis

    Talukder, S., Yue, Y., and Gkioxari, G. Totem: Tokenized time series embeddings for general time series analysis. arXiv preprint arXiv:2402.16412, 2024

    2024 arXiv

  26. [34]

    Neural discrete representation learning

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

    2017

  27. [35]

    and Breckon, T

    Wang, Q. and Breckon, T. Unsupervised domain adaptation via structured prediction based selective pseudo-labeling. In Proceedings of the AAAI conference on artificial intelligence, volume 34, pp.\ 6243--6250, 2020

    2020

  28. [36]

    Contrast everything: A hierarchical contrastive framework for medical time-series

    Wang, Y., Han, Y., Wang, H., and Zhang, X. Contrast everything: A hierarchical contrastive framework for medical time-series. Advances in Neural Information Processing Systems, 36, 2024

    2024

  29. [37]

    R., and Cook, D

    Wilson, G., Doppa, J. R., and Cook, D. J. 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, pp.\ 1768--1778, 2020

    2020

  30. [38]

    Y., Zhang, H., Pang, R., Qin, J., Ku, A., Xu, Y., Baldridge, J., and Wu, Y

    Yu, J., Li, X., Koh, J. Y., Zhang, H., Pang, R., Qin, J., Ku, A., Xu, Y., Baldridge, J., and Wu, Y. Vector-quantized image modeling with improved vqgan. arXiv preprint arXiv:2110.04627, 2021

    2021 arXiv

  31. [39]

    Domain adaptation under target and conditional shift

    Zhang, K., Sch \"o lkopf, B., Muandet, K., and Wang, Z. Domain adaptation under target and conditional shift. In International conference on machine learning, pp.\ 819--827. Pmlr, 2013

    2013

  32. [40]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.