Pith. sign in

REVIEW 1 major objections 6 minor 1 cited by

TimePro: Efficient Multivariate Long-term Time Series Forecasting with Variable- and Time-Aware Hyper-state

T0 review · 1 major / 6 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 TimePro claims that a Mamba variable-scan state, re-sampled at adaptively learned intra-variable time points, captures per-variable delays and reaches state-of-the-art or tied results on seven of eight multivariate long-term forecasting…

desk verdict TimePro's time-tune mechanism is a genuinely new idea and the forecasting results look competitive, but the advertised O(NL) complexity does not follow from the equations as written, and the paper overclaims significance without error bars. read the letter →

arxiv 2505.20774 v1 pith:444PQSBT submitted 2025-05-27 cs.LG

Xiaowen Ma , Zhenliang Ni , Shuai Xiao , Xinghao Chen This is my paper
classification cs.LG
keywords long-termtimeseriesforecastingstatespacemodelsMambamulti-delayissuehyper-statetime-tunestrategymultivariatelinearcomplexity
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

TimePro is a Mamba-based architecture for multivariate long-term forecasting that targets the multi-delay problem, where each input variable influences the target over its own time interval. Instead of transferring a plain state uniformly across variables or time points, the model scans along the variable dimension, then uses a learned per-variable offset to sample that state at adaptively selected intra-variable time points and fuses the samples into a hyper-state. The paper reports that this time-tune mechanism achieves 12 firsts and 2 seconds across 16 metrics on eight real-world benchmarks, with state-of-the-art or tied results on seven of them, while keeping computational complexity linear in both the number of variables and the series length. The contribution, as the authors present it, is evidence that learned temporal non-uniformity rather than extra capacity is what lets a linear-complexity forecaster outperform quadratic transformer baselines.

What carries the argument

The load-bearing object is the hyper-state produced by the time-tune strategy inside HyperMamba. In place of a plain state propagated uniformly, the variable-scan state $h$ is reshaped, convolved to produce a learnable offset $\delta_h$, and sampled at reference-plus-offset coordinates by linear interpolation; a linear layer then fuses $M = 9$ sampled time points into the hyper-state $h_o$, which is combined with the Mamba parameter matrix $C$ to yield the output. This construction lets the model keep fine-grained temporal structure per variate while scanning variables, and it is implemented on GPU SRAM so that the added sampling keeps the whole pipeline at $O(NL)$ complexity.

What would settle it

Train TimePro and its non-adaptive variant on a synthetic multivariate series where the target is generated from other variables with known, distinct lags, such as variable A with delay 5 and variable B with delay 40 inside the lookback window; if the learned offsets do not concentrate near the true lags, or if the non-adaptive variant matches the adaptive one in MSE, the time-tune mechanism's role in the reported gains is not supported.

Watch

Extended reading notes

Core claim

The central claim is that a Mamba state transferred only along the variable dimension can be made time-aware without losing linear cost. After the Hyper-Scan obtains the plain state $h \in \mathbb{R}^{N \times P \times D}$ for $N$ variables and $P$ patches, a convolution produces an offset $\delta_h$, the shifted coordinates $h_{\text{samp}} = h_{\text{ref}} + \delta_h$ are sampled through differentiable linear interpolation $\psi$ to give $\hat{h} \in \mathbb{R}^{N \times P \times D \times M}$, and a linear mapping fuses the $M$ sampled points into the hyper-state $h_o$. That hyper-state is multiplied by the Mamba output matrix $C$, so each variable's output is conditioned on both the cross-variable scan and locally selected time points. On eight multivariate benchmarks with lookback 96 and horizons 96, 192, 336, and 720, TimePro reports 12 firsts and 2 seconds out of 16 averaged metrics, with the largest consistent margins on the ETT family and Exchange, and the paper credits the gain to the time-tune strategy rather than to wider layers or a deeper stack.

Load-bearing premise

The central claim rests on the premise that learning, per variable, which time points should tune the variable-scan state captures real delayed dependencies rather than merely adding model capacity; if a uniform or non-adaptive state transfer with equal capacity matches TimePro's accuracy, the time-tune mechanism is not what carries the result.

Editorial extensions

If this is right

  • If the central claim is right, uniform variable-scan or time-scan state transfer leaves accuracy on the table: the adaptive time-tune variant beats its non-adaptive counterpart on Exchange and ETTh1, and full HyperMamba beats variable-only, time-only, and sequential time-and-variable scans.
  • Linear complexity in both variables and series length would let the architecture scale to much wider channel counts; the channel-scaling experiment shows TimePro's memory and inference time growing roughly linearly where iTransformer grows quadratically.
  • On the eight benchmarks, the reported results put TimePro at or near the top for horizons 96 to 720, with the largest advantages over transformer-based baselines such as iTransformer and PatchTST.
  • The design suggests accuracy can come from where the state is sampled rather than from added depth, since most datasets saturate at two to four encoder layers at a moderate feature dimension.

Reading between the lines

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

  • Beyond the paper's own claims: if the learned offsets track true physical lags, the same hyper-state could double as an interpretable per-variable delay map, something the paper's post-hoc correlation visualization only hints at.
  • Beyond the paper's own claims: because the time-tune operator is a differentiable sampler, it could be ported as a generic sample-then-fuse block into other state-space or even non-Mamba forecasting architectures.
  • Beyond the paper's own claims: a synthetic experiment with known variable-specific lags would separate the two explanations for the reported gains, namely genuine adaptive delay modeling versus added nonlinear capacity, and would show whether the learned offset positions correspond to real delay structure.
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

1 major / 6 minor

Summary. The paper proposes TimePro, a Mamba-based architecture for multivariate long-term time series forecasting. The key idea is to scan a patched multivariate series along the variable dimension with a selective state-space model and then refine the resulting hidden state by sampling it at learned per-variable time offsets (the "time-tune" step), producing a variable- and time-aware hyper-state. The manuscript reports competitive accuracy on eight standard benchmarks (12 firsts and 2 seconds out of 16 metrics in Table 2) and claims O(NL) complexity (Section 4.3, Table 1), supported by ablations of the time-tune mechanism and hyper-scan variants.

Significance. If the efficiency claim were supported, the time-tune idea—adaptively selecting time points at which to sample the variable-scan state—is a clean and plausible improvement over uniform state transfer in Mamba-based forecasters. The paper specifies the architecture in enough detail to be reimplemented, evaluates on standard benchmarks with a full results table in the appendix, includes multiple ablations, and states code availability; these are genuine strengths. However, the advertised linear-complexity guarantee does not follow from the architecture as written, and the claimed accuracy advantage over strong baselines is often within run-to-run noise, so the headline claims need revision before the paper can be accepted.

major comments (1)
  1. [Section 4.3, Eq. (8), Table 1, Appendix B.1] Please note that Section 4.3's conclusion is load-bearing for the paper's efficiency claim, and the missing implementation detail for the projections is a separate issue from the missing statistical support. Both should be resolved before publication.
minor comments (6)
  1. [Section 5.3, Table 3] The text states that the non-adaptive variant obtains MSEs of 0.360 and 0.351 on Exchange and ETTh1, respectively, but Table 3 reports the ETTh1 value as 0.451. Please correct the inconsistency.
  2. [Section 5.1] The text says experiments are conducted on 'five widely used real datasets' but then lists eight benchmarks (four ETT subsets plus Exchange, Electricity, Weather, and Solar-Energy). Please make the count consistent.
  3. [Section 5.2] The paragraph about the Weather dataset is duplicated, and 'SOTFS' appears instead of 'SOFTS'. Please clean up the prose.
  4. [Eq. (8)] The word 'Concact' should be 'Concat'.
  5. [Table 6] The full-results table is very dense, with MSE and MAE values run together; separating them into distinct columns would greatly improve readability.
  6. [Figure 6] The x-axis label 'L' for the number of encoder layers conflicts with the lookback length L used throughout the paper; consider using the symbol gamma as in Eq. (6).

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: benchmark results are external and the time-tune mechanism is validated by ablations; the O(N L) complexity proof is inconsistent but not circular.

full rationale

TimePro is an empirical architecture paper, and its central accuracy claims are measured against standard public benchmarks (ETT, Exchange, Electricity, Weather, Solar-Energy) with previously published baselines. No equation in Sections 4.1–4.2 constructs the forecast from a fitted quantity and then re-presents that quantity as a prediction: the hyper-state is formed from the input embedding by learned projections, scans, and sampled offsets (Eqs. 8–13), and the forecast is produced by a separate linear projection (Eq. 7). The Table 3 ablation compares adaptive versus non-adaptive state tuning and reports external test MSE; it is a standard architectural ablation, not a circular validation. Figure 5 is a post-hoc Pearson-correlation visualization that does not enter the training objective, so it cannot make the mechanism true by definition. There is no load-bearing self-citation or imported uniqueness theorem; the cited prior work (Mamba, PatchTST, iTransformer, SOFTS, etc.) is external. The only serious flaw is in Section 4.3: the paper states the two linear projections in Eq. 8 cost O(N P^2 D^2), and with D constant and P proportional to L this is O(N L^2), not O(N L), so the Table 1 complexity entry does not follow from the stated equations. That is a correctness/arithmetic problem in the efficiency claim, not a reduction of a prediction to its own inputs, and it does not constitute circularity. Minor per-dataset choices of D, layer count, and patch length (Appendix B.2) are ordinary hyperparameter selection rather than fitted parameters renamed as predictions. Overall circularity score: 1.

Assumptions & free parameters 5 free parameters · 5 assumptions · 1 invented entities

The paper is an empirical architecture contribution; the main borrowed-from-data quantities are per-dataset hyperparameters (D, gamma, patch length) chosen by validation performance, plus design constants (M=9, state size 1). No fitted physical constants appear. The multi-delay motivation is an assumption rather than a measured input.

free parameters (5)
  • Embedding dimension D = 32-96 (best mostly 48)
    Selected per dataset by validation performance (Fig. 6); the paper reports only the searched range, so the reported benchmark numbers depend on this choice.
  • Encoder layer count gamma = 1-4 (best 2-4)
    Selected per dataset (Fig. 6); affects both accuracy and computational cost.
  • Patch length Pl = 8-32
    Selected per dataset (Fig. 6); controls the time granularity of the state that is sampled by the time-tune mechanism.
  • Sampling point count M = 9
    Fixed constant for adaptive time-point sampling (Sec 4.2); not swept, but a free design choice that affects how much temporal context the hyper-state retains.
  • SSM state size per dimension = 1
    Set to 1 in the variable-dimension scan (Sec 4.3); this reduces complexity but also bounds the expressiveness of the plain state before time tuning.
assumptions (5)
  • standard math Zero-order hold discretization of continuous SSM parameters (Eq. 2)
    Invoked in Sec 3 to convert continuous A, B to discrete form; standard result from Gu et al. and Gupta et al.
  • domain assumption Scanning along the variable dimension yields a state that encodes inter-variable relationships
    Core design premise in Sec 4.2, Eq 8; there is no direct proof that the scan state carries the variable correlations used by the time-tune step.
  • domain assumption The multi-delay issue materially limits the baselines on these benchmarks
    Motivates the architecture in Sec 1 and is tested only indirectly in Sec 5.3 (Fig 5), not by controlled delay-injection experiments.
  • domain assumption Baseline numbers from iTransformer and S-Mamba papers are comparable to TimePro's runs
    Table 6 states results are taken from prior papers; identical splits, seeds, and hardware are assumed without re-running baselines.
  • standard math Linear interpolation is a differentiable surrogate for discrete sampling
    Used in Eq 12 following Xiong et al. 2024; standard and differentiable.
invented entities (1)
  • Variable- and time-aware hyper-state
    purpose: Internal representation formed by sampling the variable-scan state at adaptively selected time points (Eqs. 10-13), used to produce the forecast.
    Not a falsifiable external entity; its only support is downstream forecast accuracy and a post-hoc correlation visualization (Fig. 5), so it carries no independent evidence outside this model.

how reviews work

0 comments
Cite this review

Pith. "Pith review of TimePro: Efficient Multivariate Long-term Time Series Forecasting with Variable- and Time-Aware Hyper-state." pith.science (2026). https://pith.science/paper/444PQSBT

@misc{pith2026250520774,
  author       = {Pith},
  title        = {Pith review of: TimePro: Efficient Multivariate Long-term Time Series Forecasting with Variable- and Time-Aware Hyper-state},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/444PQSBT}},
  note         = {Machine review of arXiv:2505.20774}
}
read the original abstract

In long-term time series forecasting, different variables often influence the target variable over distinct time intervals, a challenge known as the multi-delay issue. Traditional models typically process all variables or time points uniformly, which limits their ability to capture complex variable relationships and obtain non-trivial time representations. To address this issue, we propose TimePro, an innovative Mamba-based model that constructs variate- and time-aware hyper-states. Unlike conventional approaches that merely transfer plain states across variable or time dimensions, TimePro preserves the fine-grained temporal features of each variate token and adaptively selects the focused time points to tune the plain state. The reconstructed hyper-state can perceive both variable relationships and salient temporal information, which helps the model make accurate forecasting. In experiments, TimePro performs competitively on eight real-world long-term forecasting benchmarks with satisfactory linear complexity. Code is available at https://github.com/xwmaxwma/TimePro.

Figures

Figures reproduced from arXiv: 2505.20774 by the authors.

Figure 1
Figure 1. Forecasting performance comparison of Time [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 3
Figure 3. Overview of our TimePro method. The multivariate time series is first embedded along the temporal dimension [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Implementation details of hardware-aware hyper-scan. We effectively apply the GPU memory hierarchy, i.e., [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figures from the paper (7 more)
Figure 5
Figure 5. Figure 5: Visualization for multivariate correlation analysis on ETTm1 (upper) and ETTh1(bottom) dataset. The visualization [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Influence of the hidden dimension of features D, Layer of encoder L and length of patches. We select ECL, [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 7
Figure 7. Figure 7: Influence of lookback window length L ∈ 48, 96, 192, 336 on ETTm2, ECL and Exchange dataset. TimePro performs almost consistently better than other models under different lookback window lengths [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 8
Figure 8. Figure 8: Comparison of forecasts between TimePro, S-Mamba and iTransformer on ETTh2 dataset when the input length [PITH_FULL_IMAGE:figures/full_fig_p014_8.png]
Figure 9
Figure 9. Figure 9: Comparison of forecasts between TimePro, S-Mamba and iTransformer on ECL dataset when the input length is [PITH_FULL_IMAGE:figures/full_fig_p015_9.png]
Figure 10
Figure 10. Figure 10: Visual comparison of different Hyper-scan designs on ETTh2 dataset when the input length is 96 and the forecast [PITH_FULL_IMAGE:figures/full_fig_p015_10.png]
Figure 11
Figure 11. Figure 11: Memory and inference time of different methods. We set the lookback window L = 96, forecast horizon H = [PITH_FULL_IMAGE:figures/full_fig_p016_11.png]

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. TSSM: Triaxial State Space Model for Global Station Weather Forecasting with Temporal-Variable-Historical Modeling

    cs.LG 2026-07 conditional novelty 6.0 of 10

    A weather forecasting model that scans time, variables, and aligned past years jointly beats prior station-forecasting baselines on Weather-5K, with particularly large reported gains on extreme-event and long-horizon metrics.

Reference graph

Works this paper leans on

38 extracted references · 19 canonical work pages · cited by 1 Pith paper

  1. [1]

    Ahamed, M. A. and Cheng, Q. Timemachine: A time series is worth 4 mambas for long-term forecasting. arXiv preprint arXiv:2403.09898, 2024

    arXiv 2024

  2. [2]

    Long-term forecasting with tide: Time-series dense encoder, 2023

    Das, A., Kong, W., Leach, A., Mathur, S., Sen, R., and Yu, R. Long-term forecasting with tide: Time-series dense encoder, 2023. URL https://arxiv.org/abs/2304.08424

    arXiv 2023

  3. [3]

    and Dao, T

    Gu, A. and Dao, T. Mamba: Linear-time sequence modeling with selective state spaces, 2024. URL https://arxiv.org/abs/2312.00752

    arXiv 2024

  4. [4]

    Combining recurrent, convolutional, and continuous-time models with linear state space layers

    Gu, A., Johnson, I., Goel, K., Saab, K., Dao, T., Rudra, A., and R\' e , C. Combining recurrent, convolutional, and continuous-time models with linear state space layers. In Ranzato, M., Beygelzimer, A., Dauphin, Y., Liang, P., and Vaughan, J. W. (eds.), Advances in Neural Information Processing Systems, volume 34, pp.\ 572--585. Curran Associates, Inc., ...

    work page 2021

  5. [5]

    Efficiently modeling long sequences with structured state spaces, 2022

    Gu, A., Goel, K., and Ré, C. Efficiently modeling long sequences with structured state spaces, 2022. URL https://arxiv.org/abs/2111.00396

    arXiv 2022

  6. [6]

    Diagonal state spaces are as effective as structured state spaces

    Gupta, A., Gu, A., and Berant, J. Diagonal state spaces are as effective as structured state spaces. In Koyejo, S., Mohamed, S., Agarwal, A., Belgrave, D., Cho, K., and Oh, A. (eds.), Advances in Neural Information Processing Systems, volume 35, pp.\ 22982--22994. Curran Associates, Inc., 2022

    work page 2022

  7. [7]

    SOFTS : Efficient multivariate time series forecasting with series-core fusion

    Han, L., Chen, X.-Y., Ye, H.-J., and Zhan, D.-C. SOFTS : Efficient multivariate time series forecasting with series-core fusion. In The Thirty-eighth Annual Conference on Neural Information Processing Systems, 2024. URL https://openreview.net/forum?id=89AUi5L1uA

    work page 2024

  8. [8]

    Recurrent neural networks for time series forecasting: Current status and future directions

    Hewamalage, H., Bergmeir, C., and Bandara, K. Recurrent neural networks for time series forecasting: Current status and future directions. International Journal of Forecasting, 37 0 (1): 0 388--427, 2021

    work page 2021

Show all 38 references
  1. [9]

    Reversible instance normalization for accurate time-series forecasting against distribution shift

    Kim, T., Kim, J., Tae, Y., Park, C., Choi, J.-H., and Choo, J. Reversible instance normalization for accurate time-series forecasting against distribution shift. In International Conference on Learning Representations, 2022. URL https://openreview.net/forum?id=cGDAkQo1C0p

    2022

  2. [10]

    Kingma, D. P. and Ba, J. Adam: A method for stochastic optimization, 2017. URL https://arxiv.org/abs/1412.6980

    2017 arXiv

  3. [11]

    Modeling long- and short-term temporal patterns with deep neural networks

    Lai, G., Chang, W.-C., Yang, Y., and Liu, H. Modeling long- and short-term temporal patterns with deep neural networks. SIGIR '18, pp.\ 95–104, New York, NY, USA, 2018. Association for Computing Machinery. ISBN 9781450356572. doi:10.1145/3209978.3210006. URL https://doi.org/10...

    2018

  4. [12]

    Bi-mamba+: Bidirectional mamba for time series forecasting

    Liang, A., Jiang, X., Sun, Y., and Lu, C. Bi-mamba+: Bidirectional mamba for time series forecasting. arXiv preprint arXiv:2404.15772, 2024

    2024 arXiv

  5. [13]

    Timemixer: Decomposable multiscale mixing for time series forecasting

    Liu, H., Zhang, J., Li, Y., Zhou, H., Zhang, S., and Xie, X. Timemixer: Decomposable multiscale mixing for time series forecasting. In International Conference on Learning Representations (ICLR), 2024 a . URL https://arxiv.org/pdf/2405.14616

    2024 arXiv

  6. [14]

    Scinet: Time series modeling and forecasting with sample convolution and interaction

    LIU, M., Zeng, A., Chen, M., Xu, Z., LAI, Q., Ma, L., and Xu, Q. Scinet: Time series modeling and forecasting with sample convolution and interaction. In Koyejo, S., Mohamed, S., Agarwal, A., Belgrave, D., Cho, K., and Oh, A. (eds.), Advances in Neural Information Processing S...

    2022

  7. [15]

    Non-stationary transformers: Exploring the stationarity in time series forecasting

    Liu, Y., Wu, H., Wang, J., and Long, M. Non-stationary transformers: Exploring the stationarity in time series forecasting. In Koyejo, S., Mohamed, S., Agarwal, A., Belgrave, D., Cho, K., and Oh, A. (eds.), Advances in Neural Information Processing Systems, volume 35, pp.\ 988...

    2022

  8. [16]

    itransformer: Inverted transformers are effective for time series forecasting

    Liu, Y., Hu, T., Zhang, H., Wu, H., Wang, S., Ma, L., and Long, M. itransformer: Inverted transformers are effective for time series forecasting. In The Twelfth International Conference on Learning Representations, 2024 b . URL https://openreview.net/forum?id=JePfAI8fah

    2024

  9. [17]

    A mamba foundation model for time series forecasting

    Ma, H., Chen, Y., Zhao, W., Yang, J., Ji, Y., Xu, X., Liu, X., Jing, H., Liu, S., and Yang, G. A mamba foundation model for time series forecasting. arXiv preprint arXiv:2411.02941, 2024

    2024 arXiv

  10. [18]

    Nguyen, N., Sinthong, P., and Kalagnanam, J

    Nie, Y., H. Nguyen, N., Sinthong, P., and Kalagnanam, J. A time series is worth 64 words: Long-term forecasting with transformers. In International Conference on Learning Representations, 2023

    2023

  11. [19]

    Ramachandran, P., Zoph, B., and Le, Q. V. Searching for activation functions, 2017. URL https://arxiv.org/abs/1710.05941

    2017 arXiv

  12. [20]

    S., Seeger, M

    Rangapuram, S. S., Seeger, M. W., Gasthaus, J., Stella, L., Wang, Y., and Januschowski, T. Deep state space models for time series forecasting. In Bengio, S., Wallach, H., Larochelle, H., Grauman, K., Cesa-Bianchi, N., and Garnett, R. (eds.), Advances in Neural Information Pro...

    2018

  13. [21]

    N., Kaiser, L., and Polosukhin, I

    Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., and Polosukhin, I. Attention is all you need, 2023. URL https://arxiv.org/abs/1706.03762

    2023 arXiv

  14. [22]

    Is mamba effective for time series forecasting? Neurocomputing, 619: 0 129178, 2025

    Wang, Z., Kong, F., Feng, S., Wang, M., Yang, X., Zhao, H., Wang, D., and Zhang, Y. Is mamba effective for time series forecasting? Neurocomputing, 619: 0 129178, 2025

    2025

  15. [23]

    Autoformer: Decomposition transformers with auto-correlation for long-term series forecasting

    Wu, H., Xu, J., Wang, J., and Long, M. Autoformer: Decomposition transformers with auto-correlation for long-term series forecasting. In Ranzato, M., Beygelzimer, A., Dauphin, Y., Liang, P., and Vaughan, J. W. (eds.), Advances in Neural Information Processing Systems, volume 3...

    2021

  16. [24]

    Timesnet: Temporal 2d-variation modeling for general time series analysis

    Wu, H., Hu, T., Liu, Y., Zhou, H., Wang, J., and Long, M. Timesnet: Temporal 2d-variation modeling for general time series analysis. In The Eleventh International Conference on Learning Representations, 2023. URL https://openreview.net/forum?id=ju_Uqw384Oq

    2023

  17. [25]

    Efficient deformable convnets: Rethinking dynamic and sparse operator for vision applications

    Xiong, Y., Li, Z., Chen, Y., Wang, F., Zhu, X., Luo, J., Wang, W., Lu, T., Li, H., Qiao, Y., Lu, L., Zhou, J., and Dai, J. Efficient deformable convnets: Rethinking dynamic and sparse operator for vision applications. In Proceedings of the IEEE/CVF Conference on Computer Visio...

    2024

  18. [26]

    Sst: Multi-scale hybrid mamba-transformer experts for long-short range time series forecasting, 2024

    Xu, X., Chen, C., Liang, Y., Huang, B., Bai, G., Zhao, L., and Shu, K. Sst: Multi-scale hybrid mamba-transformer experts for long-short range time series forecasting, 2024. URL https://arxiv.org/abs/2404.14757

    2024

  19. [27]

    Are transformers effective for time series forecasting? Proceedings of the AAAI Conference on Artificial Intelligence, 37 0 (9): 0 11121--11128, Jun

    Zeng, A., Chen, M., Zhang, L., and Xu, Q. Are transformers effective for time series forecasting? Proceedings of the AAAI Conference on Artificial Intelligence, 37 0 (9): 0 11121--11128, Jun. 2023. doi:10.1609/aaai.v37i9.26317. URL https://ojs.aaai.org/index.php/AAAI/article/v...

    2023 doi

  20. [28]

    Less is more: Fast multivariate time series forecasting with light sampling-oriented mlp structures, 2022

    Zhang, T., Zhang, Y., Cao, W., Bian, J., Yi, X., Zheng, S., and Li, J. Less is more: Fast multivariate time series forecasting with light sampling-oriented mlp structures, 2022. URL https://arxiv.org/abs/2207.01186

    2022 arXiv

  21. [29]

    and Yan, J

    Zhang, Y. and Yan, J. Crossformer: Transformer utilizing cross-dimension dependency for multivariate time series forecasting. In The Eleventh International Conference on Learning Representations, 2023. URL https://openreview.net/forum?id=vSVLM2j9eie

    2023

  22. [30]

    Informer: Beyond efficient transformer for long sequence time-series forecasting

    Zhou, H., Zhang, S., Peng, J., Zhang, S., Li, J., Xiong, H., and Zhang, W. Informer: Beyond efficient transformer for long sequence time-series forecasting. Proceedings of the AAAI Conference on Artificial Intelligence, 35 0 (12): 0 11106--11115, May 2021. doi:10.1609/aaai.v35...

    2021 doi

  23. [31]

    and Gitter, A

    Chandereng, T. and Gitter, A. Lag penalized weighted correlation for time series clustering. BMC bioinformatics, 21: 0 1--15, 2020

    2020

  24. [32]

    Linearly-evolved transformer for pan-sharpening

    Hou, J., Cao, Z., Zheng, N., Li, X., Chen, X., Liu, X., Cong, X., Hong, D., and Zhou, M. Linearly-evolved transformer for pan-sharpening. In Proceedings of the 32nd ACM International Conference on Multimedia, pp.\ 1486--1494, 2024

    2024

  25. [33]

    Hou, J., Liu, X., Wu, C., Cong, X., Huang, C., Deng, L.-J., and You, J. W. Bidomain uncertainty gated recursive network for pan-sharpening. Information Fusion, 118: 0 102938, 2025

    2025

  26. [34]

    Liu, X., Hou, J., Cong, X., Shen, H., Lou, Z., Deng, L.-J., and You, J. W. Rethinking pan-sharpening via spectral-band modulation. IEEE Transactions on Geoscience and Remote Sensing, 62: 0 1--16, 2023

    2023

  27. [35]

    Channel-aware low-rank adaptation in time series forecasting

    Nie, T., Mei, Y., Qin, G., Sun, J., and Ma, W. Channel-aware low-rank adaptation in time series forecasting. In Proceedings of the 33rd ACM International Conference on Information and Knowledge Management, pp.\ 3959--3963, 2024 a

    2024

  28. [36]

    Contextualizing mlp-mixers spatiotemporally for urban traffic data forecast at scale

    Nie, T., Qin, G., Sun, L., Ma, W., Mei, Y., and Sun, J. Contextualizing mlp-mixers spatiotemporally for urban traffic data forecast at scale. IEEE Transactions on Intelligent Transportation Systems, 2024 b

    2024

  29. [37]

    Multivariate time delay analysis based local kpca fault prognosis approach for nonlinear processes

    Xu, Y., Liu, Y., and Zhu, Q. Multivariate time delay analysis based local kpca fault prognosis approach for nonlinear processes. Chinese Journal of Chemical Engineering, 24 0 (10): 0 1413--1422, 2016. ISSN 1004-9541. doi:https://doi.org/10.1016/j.cjche.2016.06.011. URL https:/...

    2016 doi

  30. [38]

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