Pith. sign in

REVIEW 3 major objections 5 minor 39 references

On the Necessity of Multi-Domain Explanation: An Uncertainty Principle Approach for Deep Time Series Models

T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read This paper claims that the uncertainty principle gives a sufficient condition that time- and frequency-domain explanations of a deep time series model highlight different features, so presenting only the single most sparse domain can…

desk verdict A genuinely new diagnostic idea, but the sufficiency claim needs a proof that attribution pairs obey the uncertainty principle; worth refereeing. read the letter →

arxiv 2506.03267 v1 pith:2R7RLZVR submitted 2025-06-03 cs.LG cs.HC

Shahbaz Rezaei , Avishai Halev , Xin Liu This is my paper
classification cs.LGcs.HC
keywords timeseriesexplainabilityuncertaintyprinciplemulti-domainexplanationattributionmethodsfrequencydomainDonoho-Starkboundspacedeeplearninginterpretability
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 tries to establish that explaining a deep time series model in a single domain—typically time—is often not enough, because the time-domain attribution and the frequency-domain attribution can highlight genuinely different features. Its tool is the uncertainty principle: a signal cannot be sharply concentrated in both time and frequency, so if two independently generated attributions are both concentrated, the explanations cannot be describing the same feature. The paper uses the Donoho–Stark inequality as a sufficient test for this mismatch and verifies it on a synthetic dataset with known shapelet and frequency features plus several real classification and forecasting datasets. The intended practical consequence is that for many time series applications the end user should be shown both domains, not just the domain in which the attribution is most sparse. The criterion is deliberately one-directional: a violation proves different features, while no violation is inconclusive.

What carries the argument

The machinery is the Donoho–Stark uncertainty principle for discrete signals, stated with $\epsilon$-concentration: a sequence is $\epsilon$-concentrated on an index set if the energy outside that set is at most $\epsilon$. For a length-$N$ signal, the number $N_t$ of significant time coefficients and $N_f$ of significant frequency coefficients must satisfy $N_t N_f \ge N(1-(\epsilon_t+\epsilon_f))^2$. The paper feeds this inequality a pair of normalized absolute attributions—one computed in the time domain by a standard XAI method, one computed in the frequency domain by the explanation-space wrapper—and performs a grid search over threshold pairs to find the tightest concentrations. If any threshold pair violates the inequality, the algorithm declares a UP violation, meaning the two attribution maps cannot represent the same feature.

What would settle it

Take a synthetic sample with exactly one known feature, say a single shapelet or a single dominant frequency, and compute both attributions. The paper's sufficient-condition logic predicts no UP violation when the model truly uses only that one feature; a confirmed violation on such a sample shows the detector is responding to the explanation method, not to feature mismatch. A more direct test is to take the discrete Fourier transform of the time-domain attribution and compare it with the wrapper-generated frequency attribution: if the two are not approximately a Fourier pair even when the inequality $N_t N_f \ge N(1-(\epsilon_t+\epsilon_f))^2$ holds, the pair does not obey the uncertainty principle's precondition.

Watch

Extended reading notes

Core claim

The central claim is that a time-domain attribution and a frequency-domain attribution of the same sample are subject to the same localization trade-off as a function and its Fourier transform, even though the paper generates the two attributions independently rather than by transforming one into the other. If both attributions are simultaneously concentrated, the Donoho–Stark bound $N_t N_f \ge N(1-(\epsilon_t+\epsilon_f))^2$ is violated; the paper takes this violation as a sufficient condition that the two explanations emphasize different features and therefore that both explanations should be presented. The paper supports the claim by showing frequent violations across five model architectures, seven attribution methods, and datasets chosen to contain both time-localized (shapelet-like) and frequency-localized (seasonal, periodic) structure. Its own stated caveat is that the test is conservative: a violation is conclusive, but the absence of a violation leaves open whether the domains agree.

Load-bearing premise

The load-bearing premise is that an attribution pair describing the same feature in time and frequency behaves like a Fourier-transform pair, so the uncertainty inequality applies to it; the paper asserts this informally but never proves the relationship, and if it fails, a UP violation can reflect the XAI method rather than genuinely different features.

Editorial extensions

If this is right

  • For any sample where the test fires, a single-domain explanation is insufficient: the time and frequency attribution maps carry distinct information and both should be shown to the user.
  • On periodic forecasting datasets such as traffic, pedestrian counts, and electricity demand, the frequent violations imply that seasonality is visible in the frequency attribution while last-step effects are visible in the time attribution, so time-only explanations hide a class of learned features.
  • The conventional rule of presenting the most sparse domain is unsafe, because a non-localized but real feature is diluted across that domain and will not appear as a highlight.
  • Because the criterion is sufficient but not necessary, the reported violation rates understate how often multi-domain explanation is needed.
  • The behavior of the XAI method matters: methods with baseline or sparsity mechanisms (such as LIME) can create violations on their own, while a baseline-free method such as Saliency never violated the bound in these experiments.

Reading between the lines

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

  • Beyond the paper: the whole sufficient-condition reading rests on an unproved premise—that when a time attribution and a frequency attribution describe the same feature, the pair behaves like a Fourier-transform pair. If an XAI method's sparsity or baseline effect creates localization on its own, a violation could be a method artifact rather than evidence of distinct features.
  • Beyond the paper: the same concentration test could be applied to other conjugate explanation pairs, such as wavelet scales or trend-seasonal decompositions, turning the test into a general 'when is multi-domain explanation needed' detector.
  • Beyond the paper: the model frequency-response plots suggest a predictive extension—probe a trained model with controlled sinusoids of varying amplitude and frequency, and use the measured response to anticipate which model-XAI combinations will produce UP violations before running any explainer.
  • Beyond the paper: the paper's own Table I reports violations on single-feature synthetic samples (for example InceptionTime with DeepLIFT on shapelet-only samples). Whether those are genuine model responses to non-features or artifacts of the attribution method is the empirical question that would separate the tool from the conclusion.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper argues that presenting only a single explanation domain (time or frequency) is insufficient for deep time series models. It proposes to use the Donoho-Stark uncertainty principle (Theorem 2, Eq. (4)) as a sufficient condition: if the time-domain and frequency-domain attributions violate the bound Nt*Nf >= N(1-(eps_t+eps_f))^2, then the features they highlight cannot be the same, so both domains should be presented. The frequency attributions are obtained via the authors' explanation-space wrapper. The method is evaluated with seven XAI methods on five architectures and on synthetic and real datasets, reporting frequent UP violations across many settings.

Significance. If the central sufficiency claim were correct, this would be a practically useful and theoretically grounded diagnostic for multi-domain explanation, and the paper provides a broad empirical survey of a relevant phenomenon: time- and frequency-domain attributions can be simultaneously localized. The qualitative examples (e.g., StarLightCurves, MIMIC) are instructive, and the LIME aggregation idea is a useful practical note. However, the paper's main contribution as stated is a sufficient condition, and that condition is unsupported and in fact contradicted by the paper's own mathematical setup; the empirical results cannot compensate for this because they inherit the same ambiguity. No code is released, and the frequency-attribution framework is taken from the authors' prior work [2], which is not independently validated here.

major comments (3)
  1. [Section III.A, Theorem 2 / Eq. (4)] The claim that a UP violation is a sufficient condition for the time- and frequency-domain attributions to represent different features is unproven and, as stated, false. The Donoho-Stark bound applies to a signal and its DFT, not to arbitrary pairs of attributions. The paper explicitly states that the attributions are generated independently and are 'not Fourier counterparts of each other.' For the wrapper model W(f)=M(FFT^{-1} f), attribution pairs computed by gradient-based methods are generally not DFT pairs even when the model uses a single well-defined feature. For a linear model M(x)=w·x, InputXGradient gives A_t = x⊙w in time and A_f = f⊙(F w) in frequency, whereas DFT(A_t) = (F x) * (F w) (a convolution) rather than A_f. These two attributions can be simultaneously sparse for a single-feature model (e.g., a shapelet weight w and a narrowband input x), yielding a UP violation that is an artifact of the attribution method, not evidence of two distinct model features. Thus the paper's sufficiency assertion, and the statement in Section V that the approach 'essentially gives no false positive,' are not justified.
  2. [Section IV-A, Tables I and II, and Section IV-B] The post-hoc exclusion of LIME results weakens the empirical claim. The paper first reports very high violation rates for LIME in Tables I and II and then says 'We ignore the results of LIME here' because its sparsity regularization artificially induces UP violations. This directly illustrates that UP violations can be method artifacts, not indicators of genuinely distinct features, and the exclusion criterion is introduced only after seeing the results. A pre-registered or otherwise justified exclusion rule would be needed to avoid the appearance of cherry-picking. Moreover, the high LIME violation rate is consistent with the concern raised by the unproven Fourier-pair assumption: any method that makes both attributions sparse will violate the bound regardless of the underlying model features.
  3. [Section IV-C and Table I] The synthetic ground-truth experiments do not validate the sufficiency condition. Table I shows UP violations on samples that contain only a time feature or only a frequency feature (e.g., DeepLift with InceptionTime: 42% on time-only class 0). The authors argue in Section IV-C that the model's frequency response makes it genuinely responsive to non-features, so these are not false positives. This explanation may be correct, but it means that a UP violation does not tell us whether the attributions highlight features a human would regard as distinct; it can also reflect spurious sensitivity of the specific model/XAI combination. Without a controlled setting in which the model is known to use exactly one feature and the XAI method is verified to be faithful to that feature, the experiment cannot establish that UP violations are a reliable indicator of the need for multi-domain explanation.
minor comments (5)
  1. [Section III.A] The explanation-space wrapper is never formally defined; please specify the wrapper model W(f)=M(FFT^{-1} f) and the exact attribution procedure used on the wrapper, since the paper's technical claims depend on this construction.
  2. [Section III.B, Algorithm 1] The relationship between the sorted thresholds tsteps[i] and fsteps[j] and the resulting Nt, Nf values is described informally; a step-by-step derivation of why this grid search finds the tightest (epsilon_t, epsilon_f) pair for each (Nt, Nf) would improve clarity.
  3. [Section IV-A] There are several typos: 'UP voilation' should be 'UP violation'; 'stem form' appears to be a misspelling of 'stem from'; in the Figure 1 caption, 'wrapped method' should be 'wrapper method.'
  4. [Section V] The statement 'which essentially gives no false positive' is a claim that depends entirely on the unproven sufficiency property; it should be either proven or removed, given the LIME counterexample and the single-feature violations in Table I.
  5. [Section IV-A] The discussion of MIMICPerformance not being z-normalized is interesting but the connection to UP violations would be clearer if the authors explained how the model exploits the mean value and why that manifests as a frequency-domain feature.

Circularity Check

0 steps flagged · score 2.0 of 10

The UP criterion itself is not circular: it applies an external Donoho-Stark bound and no fitted parameter is renamed as a prediction. The only self-citation dependency is the authors' own explanation-space wrapper [2], used as a tool to generate frequency attributions; this is a methodological reliance rather than a reduction of the central claim.

full rationale

The paper's detection rule is Theorem 2 of [18] applied to two attribution vectors. Nothing is fitted to a subset of data and then reported as a prediction: Algorithm 1 is a deterministic grid search over concentration thresholds. The central inference, that a UP violation is sufficient for the time and frequency attributions not to represent the same feature, is an unproven conceptual bridge rather than a circular reduction. The paper explicitly states that the attributions are generated independently and are not Fourier counterparts, so the Donoho-Stark inequality does not apply to them by construction, and no lemma is supplied showing that a faithfully reported single feature would yield a Fourier-pair attribution. That is a correctness risk, not an equation that reduces to its inputs. The main self-citation is reference [2], the authors' prior explanation-space wrapper, which is used throughout to generate frequency attributions and for which no code or independent validation is provided. This is a minor self-citation concern, but the wrapper is not a uniqueness theorem or the sole possible route: other frequency-domain attribution methods such as DFT-LPR [3] could supply the same input to the UP test. The central claim therefore retains independent content, and the circularity score is low.

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

No fitted constants or invented entities are introduced. The central claim rests on the standard uncertainty theorem, the unproven Fourier-pair assumption about attributions, and the validity of the self-cited explanation space wrapper.

assumptions (3)
  • standard math The Donoho-Stark uncertainty principle (Theorem 2) is valid for finite sequences and their discrete Fourier transforms.
    External result from [18]; the paper relies on it in Eq. (4) and Algorithm 1.
  • ad hoc to paper If time and frequency attributions represent the same feature, they must behave like a Fourier-transform pair and therefore cannot both be sharply localized.
    This is the bridge from 'UP violated' to 'features differ'. Section III.A asserts this informally but does not prove it, and it is load-bearing for the paper's main claim.
  • domain assumption The explanation space wrapper from reference [2] produces faithful frequency-domain attributions for models trained on the time domain.
    All frequency attributions in the experiments come from this wrapper, which is the authors' own prior work. No independent validation or code is provided in this paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of On the Necessity of Multi-Domain Explanation: An Uncertainty Principle Approach for Deep Time Series Models." pith.science (2026). https://pith.science/paper/2R7RLZVR

@misc{pith2026250603267,
  author       = {Pith},
  title        = {Pith review of: On the Necessity of Multi-Domain Explanation: An Uncertainty Principle Approach for Deep Time Series Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2R7RLZVR}},
  note         = {Machine review of arXiv:2506.03267}
}
read the original abstract

A prevailing approach to explain time series models is to generate attribution in time domain. A recent development in time series XAI is the concept of explanation spaces, where any model trained in the time domain can be interpreted with any existing XAI method in alternative domains, such as frequency. The prevailing approach is to present XAI attributions either in the time domain or in the domain where the attribution is most sparse. In this paper, we demonstrate that in certain cases, XAI methods can generate attributions that highlight fundamentally different features in the time and frequency domains that are not direct counterparts of one another. This suggests that both domains' attributions should be presented to achieve a more comprehensive interpretation. Thus it shows the necessity of multi-domain explanation. To quantify when such cases arise, we introduce the uncertainty principle (UP), originally developed in quantum mechanics and later studied in harmonic analysis and signal processing, to the XAI literature. This principle establishes a lower bound on how much a signal can be simultaneously localized in both the time and frequency domains. By leveraging this concept, we assess whether attributions in the time and frequency domains violate this bound, indicating that they emphasize distinct features. In other words, UP provides a sufficient condition that the time and frequency domain explanations do not match and, hence, should be both presented to the end user. We validate the effectiveness of this approach across various deep learning models, XAI methods, and a wide range of classification and forecasting datasets. The frequent occurrence of UP violations across various datasets and XAI methods highlights the limitations of existing approaches that focus solely on time-domain explanations. This underscores the need for multi-domain explanations as a new paradigm.

Figures

Figures reproduced from arXiv: 2506.03267 by the authors.

Figure 1
Figure 1. Synthetic dataset containing both/either shapelet fea [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. UP violation on real datasets: The first two columns and the last two columns are time frequency pair of a same sample. [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. The result of removing the negatively highlighted [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (2 more)
Figure 5
Figure 5. Figure 5: InceptionTime trained on the synthetic dataset and the [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: Frequency/Amplitude response of various models [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

39 extracted references · 26 canonical work pages

  1. [2]

    Explanation Space: A New Perspective into Time Series Interpretability

    S. Rezaei and X. Liu, “Explanation space: A new perspective into time series interpretability,” arXiv preprint arXiv:2409.01354 , 2024

    work page Pith review arXiv 2024

  2. [1]

    Axiomatic attribution for deep networks,

    M. Sundararajan, A. Taly, and Q. Yan, “Axiomatic attribution for deep networks,” in International conference on machine learning . PMLR, 2017, pp. 3319–3328

    2017

  3. [3]

    Explainable ai for time series via virtual inspection layers,

    J. Vielhaben, S. Lapuschkin, G. Montavon, and W. Samek, “Explainable ai for time series via virtual inspection layers,” Pattern Recognition, vol. 150, p. 110309, 2024. (a) Synthetic - FCN (b) Synthetic - TST (c) Synthetic - ResNet (d) Synthetic - InceptionTime (e) FordA - TCN (f) MIMICPerformECG - ResNet Fig. 6: Frequency/Amplitude response of various mod...

    work page 2024

  4. [4]

    ¨Uber den anschaulichen inhalt der quantentheoretischen kinematik und mechanik,

    W. Heisenberg, “ ¨Uber den anschaulichen inhalt der quantentheoretischen kinematik und mechanik,” Zeitschrift f¨ur Physik, vol. 43, no. 3, pp. 172– 198, 1927

    work page 1927

  5. [5]

    The uncertainty principle: a mathematical survey,

    G. B. Folland and A. Sitaram, “The uncertainty principle: a mathematical survey,” Journal of Fourier analysis and applications , vol. 3, pp. 207– 238, 1997

    work page 1997

  6. [6]

    Learning explainable models using attribution priors.(2019),

    G. Erion, D. J. Joseph, P. Sturmfels, S. M. Lundberg, and S.-I. Lee, “Learning explainable models using attribution priors.(2019),” CoRR, abs/1906.10670, 2020

    arXiv 2019

  7. [7]

    Learning important features through propagating activation differences,

    A. Shrikumar, P. Greenside, and A. Kundaje, “Learning important features through propagating activation differences,” in International conference on machine learning . PMlR, 2017, pp. 3145–3153

    work page 2017

  8. [8]

    Deep inside convolutional networks: Visualising image classification models and saliency maps,

    K. Simonyan, A. Vedaldi, and A. Zisserman, “Deep inside convolutional networks: Visualising image classification models and saliency maps,” arXiv preprint arXiv:1312.6034 , 2013

    arXiv 2013

Show all 39 references
  1. [9]

    A unified approach to interpreting model predictions,

    S. M. Lundberg and S.-I. Lee, “A unified approach to interpreting model predictions,” Advances in neural information processing systems, vol. 30, 2017

    2017

  2. [10]

    Not just a black box: Learning important features through propagating activation differences,

    A. Shrikumar, P. Greenside, A. Shcherbina, and A. Kundaje, “Not just a black box: Learning important features through propagating activation differences,” arXiv preprint arXiv:1605.01713 , 2016

    2016 arXiv

  3. [11]

    Visualizing and understanding convo- lutional networks,

    M. D. Zeiler and R. Fergus, “Visualizing and understanding convo- lutional networks,” in Computer Vision–ECCV 2014: 13th European Conference, Zurich, Switzerland, September 6-12, 2014, Proceedings, Part I 13. Springer, 2014, pp. 818–833

    2014

  4. [12]

    Clinical intervention prediction and understanding using deep networks,

    H. Suresh, N. Hunt, A. Johnson, L. A. Celi, P. Szolovits, and M. Ghas- semi, “Clinical intervention prediction and understanding using deep networks,” arXiv preprint arXiv:1705.08498 , 2017

    2017 arXiv

  5. [13]

    Benchmark- ing deep learning interpretability in time series predictions,

    A. A. Ismail, M. Gunady, H. Corrada Bravo, and S. Feizi, “Benchmark- ing deep learning interpretability in time series predictions,” Advances in neural information processing systems, vol. 33, pp. 6441–6452, 2020

    2020

  6. [14]

    Class-specific explainability for deep time series classifiers,

    R. Doddaiah, P. Parvatharaju, E. Rundensteiner, and T. Hartvigsen, “Class-specific explainability for deep time series classifiers,” in 2022 IEEE International conference on data mining (ICDM) . IEEE, 2022, pp. 101–110

    2022

  7. [15]

    Encoding time-series explanations through self-supervised model behavior consistency,

    O. Queen, T. Hartvigsen, T. Koker, H. He, T. Tsiligkaridis, and M. Zit- nik, “Encoding time-series explanations through self-supervised model behavior consistency,” Advances in Neural Information Processing Sys- tems, vol. 36, 2024

    2024

  8. [16]

    Agnostic local ex- planation for time series classification,

    M. Guillem ´e, V . Masson, L. Roz´e, and A. Termier, “Agnostic local ex- planation for time series classification,” in 2019 IEEE 31st international conference on tools with artificial intelligence (ICTAI) . IEEE, 2019, pp. 432–439

    2019

  9. [17]

    Inherently interpretable time series classification via multiple instance learning,

    J. Early, G. K. Cheung, K. Cutajar, H. Xie, J. Kandola, and N. Twomey, “Inherently interpretable time series classification via multiple instance learning,” arXiv preprint arXiv:2311.10049 , 2023

    2023 arXiv

  10. [18]

    Uncertainty principles and signal recovery,

    D. L. Donoho and P. B. Stark, “Uncertainty principles and signal recovery,” SIAM Journal on Applied Mathematics , vol. 49, no. 3, pp. 906–931, 1989

    1989

  11. [19]

    Entropic uncertainty relations and their applications,

    P. J. Coles, M. Berta, M. Tomamichel, and S. Wehner, “Entropic uncertainty relations and their applications,” Reviews of Modern Physics, vol. 89, no. 1, p. 015002, 2017

    2017

  12. [20]

    The uncertainty principle: variations on a theme,

    A. Wigderson and Y . Wigderson, “The uncertainty principle: variations on a theme,” Bulletin of the American Mathematical Society , vol. 58, no. 2, pp. 225–261, 2021

    2021

  13. [21]

    Fourier transforms of functions supported on sets of finite lebesgue measure

    M. Benedicks, “Fourier transforms of functions supported on sets of finite lebesgue measure.” J. MATH. ANAL. APPLIC. , vol. 106, no. 1, pp. 180–183, 1985

    1985

  14. [22]

    A theorem concerning fourier transforms,

    G. Hardy, “A theorem concerning fourier transforms,” Journal of the London Mathematical Society , vol. 1, no. 3, pp. 227–231, 1933

    1933

  15. [23]

    Inequalities in fourier analysis,

    W. Beckner, “Inequalities in fourier analysis,” Annals of Mathematics , vol. 102, no. 1, pp. 159–182, 1975

    1975

  16. [24]

    Inceptiontime: Finding alexnet for time series classification,

    H. Ismail Fawaz, B. Lucas, G. Forestier, C. Pelletier, D. F. Schmidt, J. Weber, G. I. Webb, L. Idoumghar, P.-A. Muller, and F. Petitjean, “Inceptiontime: Finding alexnet for time series classification,” Data Mining and Knowledge Discovery , vol. 34, no. 6, pp. 1936–1962, 2020

    1936

  17. [25]

    An empirical evaluation of generic convolutional and recurrent networks for sequence modeling,

    S. Bai, J. Z. Kolter, and V . Koltun, “An empirical evaluation of generic convolutional and recurrent networks for sequence modeling,” arXiv preprint arXiv:1803.01271, 2018

    2018 arXiv

  18. [26]

    A transformer-based framework for multivariate time series representation learning,

    G. Zerveas, S. Jayaraman, D. Patel, A. Bhamidipaty, and C. Eickhoff, “A transformer-based framework for multivariate time series representation learning,” in Proceedings of the 27th ACM SIGKDD conference on knowledge discovery & data mining , 2021, pp. 2114–2124

    2021

  19. [27]

    tsai - a state-of-the-art deep learning library for time series and sequential data,

    I. Oguiza, “tsai - a state-of-the-art deep learning library for time series and sequential data,” Github, 2023. [Online]. Available: https://github.com/timeseriesAI/tsai

    2023

  20. [28]

    ” why should i trust you?

    M. T. Ribeiro, S. Singh, and C. Guestrin, “” why should i trust you?” explaining the predictions of any classifier,” in Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining, 2016, pp. 1135–1144

    2016

  21. [29]

    Cap- tum: A unified and generic model interpretability library for pytorch,

    N. Kokhlikyan, V . Miglani, M. Martin, E. Wang, B. Alsallakh, J. Reynolds, A. Melnikov, N. Kliushkina, C. Araya, S. Yan et al., “Cap- tum: A unified and generic model interpretability library for pytorch,” arXiv preprint arXiv:2009.07896 , 2020

    2009 arXiv

  22. [30]

    The ucr time series archive,

    H. A. Dau, A. Bagnall, K. Kamgar, C.-C. M. Yeh, Y . Zhu, S. Gharghabi, C. A. Ratanamahatana, and E. Keogh, “The ucr time series archive,” IEEE/CAA Journal of Automatica Sinica , vol. 6, no. 6, pp. 1293–1305, 2019

    2019

  23. [31]

    MIMIC performance dataset,

    P. H. C. et al., “MIMIC performance dataset,” https://ppg- beats.readthedocs.io/en/latest/datasets/mimic perform training/, 2022

    2022

  24. [32]

    US hourly climate dataset,

    N. C. of Environmental Information, “US hourly climate dataset,” https://www.ncei.noaa.gov/metadata/geoportal/, 2020

    2020

  25. [33]

    Dhaka Stock Exchange Historical Data,

    M. S. A. Tashreef Muhammad, Rakibul Islam, “Dhaka Stock Exchange Historical Data,” https://data.mendeley.com/datasets/23553sm4tn/3, 2024

    2024

  26. [34]

    SKForecast Repository,

    J. E. O. Joaquin Amat Rodrigo, “SKForecast Repository,” https://github.com/skforecast/skforecast-datasets, 2023

    2023

  27. [35]

    Monash time series forecasting archive,

    R. Godahewa, C. Bergmeir, G. I. Webb, R. J. Hyndman, and P. Montero- Manso, “Monash time series forecasting archive,” in Neural Information Processing Systems Track on Datasets and Benchmarks , 2021

    2021

  28. [36]

    Don’t get me wrong: How to apply deep visual interpre- tations to time series,

    C. L ¨offler, W.-C. Lai, B. Eskofier, D. Zanca, L. Schmidt, and C. Mutschler, “Don’t get me wrong: How to apply deep visual interpre- tations to time series,” arXiv preprint arXiv:2203.07861 , 2022

    2022 arXiv

  29. [37]

    Finding anomalous periodic time series: An application to catalogs of periodic variable stars,

    U. Rebbapragada, P. Protopapas, C. E. Brodley, and C. Alcock, “Finding anomalous periodic time series: An application to catalogs of periodic variable stars,” Machine learning, vol. 74, pp. 281–313, 2009

    2009

  30. [38]

    The secret lives of cepheids: A multi-wavelength study of the atmospheres and real-time evolution of classical cepheids,

    S. G. Engle, “The secret lives of cepheids: A multi-wavelength study of the atmospheres and real-time evolution of classical cepheids,” arXiv preprint arXiv:1504.02713, 2015

    2015 arXiv

  31. [39]

    Kallrath, E

    J. Kallrath, E. F. Milone, and R. Wilson, Eclipsing binary stars: modeling and analysis . Springer, 2009, vol. 11

    2009

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.