BEAT: Balanced Frequency Adaptive Tuning for Long-Term Time-Series Forecasting

Naipeng Chen; Sanghoon Lee; Seonghwa Choi; Weisi Lin; Zhixuan Li

REVIEW 3 major objections 4 minor 47 references

BEAT: Balanced Frequency Adaptive Tuning for Long-Term Time-Series Forecasting

T0 review · 3 major / 4 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read BEAT, a training-only gradient-balancing framework for wavelet frequency bands, reports the most first-place results across seven long-term forecasting benchmarks.

desk verdict A plausible training-time idea undone by a self-contradictory results table and a gradient rule that does not match its stated intent. read the letter →

arxiv 2501.19065 v2 pith:WSMDS23C submitted 2025-01-31 cs.LG cs.AI

Zhixuan Li , Naipeng Chen , Seonghwa Choi , Sanghoon Lee , Weisi Lin This is my paper · ORCID

classification cs.LGcs.AI

keywords long-termtime-seriesforecastingfrequency-domainmethodswaveletdecompositiongradientmodulationbalancedtrainingconvergencemonitoringadaptivetuningmultivariate

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

BEAT (Balanced frEquency Adaptive Tuning) is a training-time framework for frequency-domain long-term time-series forecasting. It claims that training every wavelet frequency branch under one unified loss is inefficient because high-frequency branches converge too fast and low-frequency branches too slowly, so the model overfits one regime while underfitting the other. BEAT monitors each frequency's prediction error relative to the others, converts that discrepancy into a per-branch gradient modulation coefficient, and rescales back-propagated gradients to synchronize learning. If the claim is right, any wavelet-based forecaster can be upgraded at training time, with inference unchanged, and consistently beat state-of-the-art methods on long-horizon benchmarks.

What carries the argument

The load-bearing mechanism is the pair formed by the Frequency-Specific Monitor and the Dynamical Gradient Balancer. The monitor uses wavelet decomposition of the ground-truth target to compute a normalized per-frequency discrepancy ratio $r_v$ (Eq. 5), which is read as a relative learning-speed signal for each branch. The balancer converts $r_v$ into a gradient modulation coefficient $c_v$ via Eq. 7 and applies it in the parameter update as $\bar{g}_v = c_v g_v$ (Eq. 8). The design intent is that branches with small errors relative to the mean (fast learners) receive damped gradients while branches with large errors (slow learners) receive boosted gradients, bringing all frequency branches into synchronized convergence.

What would settle it

Run BEAT with all modulation coefficients fixed to a constant $c_v = 1$ while keeping the Frequency-Specific Monitor active; if accuracy does not drop, the monitoring signal alone, rather than the gradient balancing, accounts for the gains. A direct check of Eq. 7 also shows that no positive discrepancy ratio $r_v$ produces $c_v < 1$, so the claimed 'moderating gradients for rapid learners' branch is never active, which would indicate that the stated mechanism does not match the implementation.

Watch

Extended reading notes

Core claim

The central claim is that asynchronous learning across wavelet frequency bands is a measurable, correctable cause of suboptimal long-term forecasts, and that correcting it through training-time gradient modulation yields consistent gains. BEAT decomposes the ground-truth target $Y$ into an approximation coefficient $Y_A$ and detail coefficients $Y_{D_i}$ via discrete wavelet transform, computes per-branch MSE discrepancies $\delta_A$ and $\delta_{D_i}$, takes the mean $\mu$ over the detail coefficients, and defines discrepancy ratios $r_v = \delta_v / \mu$ (Eq. 5). The Dynamical Gradient Balancer then multiplies each branch's back-propagated gradient by a coefficient $c_v$ from Eq. 7, with the stated intention of damping branches that have learned quickly and boosting branches that lag. In experiments on seven datasets with prediction horizons $\{96, 192, 336, 720\}$, BEAT reports the highest number of first-place MSE and MAE results among WPMixer, TimeMixer, iTransformer, TSMixer, PatchTST, Crossformer, TiDE, and DLinear.

Load-bearing premise

The whole method depends on one assumption: that each frequency band's forecast error relative to the average reliably marks it as a fast or slow learner, and that the gradient-scaling rule then dampens fast learners and boosts slow learners exactly as intended.

Editorial extensions

If this is right

Frequency-decomposition forecasters trained with BEAT should exhibit more synchronized convergence across wavelet bands, reducing overfitting of high-frequency branches and underfitting of low-frequency branches.
Because BEAT changes only the training loop, the inference-time cost of the base forecaster is unchanged.
The per-frequency discrepancy ratio gives a real-time, training-set-based signal for convergence, overfitting, and underfitting without requiring extra validation data.
On the seven reported benchmarks, BEAT attains more first-place MSE and MAE results than nine state-of-the-art baselines across prediction lengths $\{96, 192, 336, 720\}$.

Reading between the lines

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

Because BEAT is training-only and does not alter the base architecture, the same discrepancy-ratio monitor could in principle be attached to Fourier-based or other multiband forecasters, not just wavelet-based WPMixer.
The monitoring signal uses ground-truth wavelet coefficients only during training, which suggests a natural curriculum: a branch could be frozen or detached once its discrepancy ratio stabilizes, saving compute without changing the architecture.
The reported sensitivity of the best wavelet type and decomposition level to the dataset hints that the monitor itself could be used to select those hyperparameters automatically, something the paper does not explore.
Benchmark reporting that shows only aggregate MSE and MAE may hide frequency-level imbalance; reporting per-frequency losses would make it easier to see whether BEAT's gains come from the low-frequency branch, the high-frequency branch, or both.

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.

Desk Editor's Note

read the letter

The core idea here is legitimate: apply gradient balancing from multimodal learning to the frequency sub-networks of a wavelet-based forecaster, so that fast-converging components don't overfit while slow ones underfit. That is a new application and a sensible direction. The paper does a reasonable job motivating the problem, and the ablations on wavelet type and decomposition level are the kind of systematic check you want to see.

But the paper as written does not hold together. The central claim is that BEAT "consistently outperforms state-of-the-art approaches" and has the most first-place rankings, and the paper's own Table 3 contradicts that. The 1st Count row says 28 MSE and 30 MAE firsts, but there are only 28 cells per metric (7 datasets × 4 horizons). 30 is impossible. And it isn't a small bookkeeping error: on Traffic, iTransformer beats BEAT on both metrics at every prediction length, and on ECL 96 and Weather 336 other baselines also win. So the headline empirical claim is false in the paper's own numbers.

The mechanism has a similar problem. Equation (7) is supposed to suppress gradients for fast learners and amplify them for slow learners, but as written it always outputs a coefficient ≥ 1. The r_v ≤ 1 branch gives 1/r_v ≥ 1, and the r_v > 1 branch gives a sigmoid shifted by +0.5, which is also > 1. Nothing is ever suppressed. This is not a minor typo: the explanation of the method in the text and the actual formula describe different behaviors. If the formula is what was run, the method is not what the paper says it is.

What else weakens the case: no code is released, there are no error bars or significance tests, and the improvements over WPMixer are tiny (often 0.001–0.003 MSE), easily within run-to-run noise on these benchmarks. The ablation tables show the same small gaps. So even if the mechanism were fixed, the evidence that BEAT helps is thin.

I'd sum it up as a promising idea attached to a manuscript that currently doesn't support its own conclusions. The authors could fix the equation, correct the count, and rerun with seeds, and then the paper might be a reasonable workshop contribution or a solid component of a larger study. But in its present form, it isn't there. I would not send it to a serious referee; I'd desk-reject and invite a corrected resubmission.

Referee Report

3 major / 4 minor

Summary. This manuscript proposes BEAT, a training-time gradient modulation scheme for frequency-decomposed time-series forecasting. The method decomposes the target series via multi-level discrete wavelet transform into approximation and detail coefficients, trains a separate WPMixer-style network per coefficient, monitors per-frequency prediction-error ratios, and scales each network's gradient by a coefficient c_v defined in Eq. (7). The authors evaluate on seven datasets (Weather, four ETT variants, Traffic, ECL) for horizons 96/192/336/720 and report MSE/MAE against eight baselines, plus ablations over wavelet type, decomposition level, discrepancy metric, and loss- versus gradient-modulation. The headline claim is that BEAT 'consistently outperforms state-of-the-art approaches' and achieves the highest number of first-place rankings across all prediction lengths and datasets.

Significance. If the mechanism and results were as stated, BEAT would be a simple, inference-free improvement applicable to any frequency-decomposed forecaster, and the study would be a useful contribution to training dynamics for multi-scale forecasting. The paper covers standard benchmarks and includes ablations of several design choices, which is a strength. However, two load-bearing issues undermine the manuscript: Eq. (7) does not implement the described suppression of fast learners, and Table 3's first-place counts are arithmetically impossible and contradicted by the table's own cell values. The central empirical claim is therefore not currently supported.

major comments (3)

[Methodology, Dynamical Gradient Balancer] Equation (7) contradicts the stated intent. For r_v <= 1 the formula gives c_v = 1/r_v >= 1, and for r_v > 1 the sigmoid term is greater than 0.5 plus the offset 0.5, so c_v > 1. Thus the coefficient never falls below 1 and never 'moderates' a fast learner; it amplifies the gradients of all frequency networks, with the largest amplification assigned to the fastest learners (smallest r_v). The prose claims the opposite ('For a network with a faster learning process ... gradients will be decreased'). The formula needs to be corrected, and the prose condition 'where rv > 0' appears to be a typo for 'rv <= 1', but even that branch still amplifies.
[Experiments, Table 3] The '1st Count' row reports 28 MSE and 30 MAE first places for BEAT. Since Table 3 has only 7 datasets x 4 prediction lengths = 28 cells per metric, 30 is impossible; moreover, BEAT does not win all cells. For example, Traffic at horizons 96, 192, 336, and 720 is better for iTransformer (e.g., horizon 96: 0.395/0.268 versus BEAT 0.459/0.278; horizon 720: 0.467/0.302 versus 0.510/0.312); Weather at horizon 336 is better for TimeMixer (0.251 versus 0.258); and ECL at horizon 96 is better for iTransformer (0.148 versus 0.149). These internal inconsistencies invalidate the abstract and Section 4.1 claims of consistent superiority.
[Section 4.1, Table 3] The comparison reports single-run point estimates without standard deviations, error bars, or significance tests, although many differences are at the third decimal (e.g., ECL horizon 96 MSE 0.149 versus 0.148). Given that the paper's central claim is a ranking claim over baselines, the results need multiple seeds and a significance check; otherwise 'first-place' outcomes may reflect noise rather than a genuine advantage.

minor comments (4)

[Ablation Study, Effect of BEAT] The table reference appears as 'Table??' and should be Table 2.
[Experimental Settings, Baselines] The text says 'nine state-of-the-art baseline methods' but lists eight methods (WPMixer, TimeMixer, iTransformer, TSMixer, PatchTST, Crossformer, TiDE, DLinear).
[Methodology, Equation (8)] The notation \bar{g}L(\theta_v^u) is overloaded; \bar g_v should be defined as c_v g_v, and the prose 'Gradient Decent' should be 'Gradient Descent'.
[Methodology, Equation (1)] The output order in Eq. (1) lists YD_f, YD_{f-1}, ..., YD_1 while the text later refers to i in {1, ..., f}; the ordering convention should be stated once to avoid ambiguity about which index corresponds to the highest frequency.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: BEAT's training-time gradient balancing is not equivalent to its inputs; Table 3 inconsistencies are correctness issues, not circularity.

full rationale

BEAT's derivation chain is a training-time feedback loop: per-frequency discrepancies (Eqs. 2-5) are computed from the current batch's ground-truth decomposition and predictions, and the resulting ratios modulate back-propagated gradients (Eqs. 7-8). The test target never enters the modulation-coefficient computation, so the reported forecasts are not equal to any fitted quantity by construction. The only self-citations are generic related-work pointers (Li et al. 2022, 2023) and are not load-bearing. Hyperparameters such as wavelet type and decomposition level are ablated on reported test metrics (Section 4.3), which is a selection-on-test-set methodological concern, but the resulting advantage is not a forced reduction: the ablation differences are small and BEAT still outperforms most baseline cells even under alternative choices, so this does not rise to circularity. The internal inconsistency in Table 3 (e.g., Traffic losses to iTransformer; the '1st Count' of 30 MAE firsts exceeding the 28 cells per metric available) is a factual/correctness problem, not a circular-derivation problem. Overall, the central method and evaluation are self-contained against external benchmarks, with no significant circularity.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The method introduces algorithmic components (FSM, DGB) but no new physical or mathematical entities; the key load is carried by the modulation formula in Eq. 7, whose behavior is not validated.

free parameters (3)

Wavelet decomposition level f = 2 (Weather), 3 (ETTh1), varies
Chosen per dataset based on validation performance; affects model capacity and the number of gradient-modulated branches.
Wavelet type ψ = Daubechies (Weather), Coiflets (ETTh1)
Selected by validation; Table 4 shows the best type differs by dataset.
Gradient modulation temperature (0.5) and offset (0.5) = 0.5, 0.5
Hand-designed constants in Eq. 7; no sensitivity analysis is provided.

assumptions (4)

standard math Multi-level discrete wavelet transform is invertible and the decomposition preserves all information needed for forecasting.
Invoked in Methodology when decomposing X and Y into approximation and detail coefficients.
domain assumption The per-frequency discrepancy ratio r_v, normalized by the mean of detail discrepancies, is a reliable indicator of per-frequency learning speed.
This assumption is the basis of FSM; no empirical validation is provided beyond the final accuracy numbers.
ad hoc to paper Scaling each frequency-specific network's gradient by c_v as defined in Eq. 7 synchronizes learning across frequencies.
The formula is chosen without derivation, and its printed form contradicts the stated intent of suppressing fast learners.
domain assumption Per-network gradient scaling does not interfere with the overall optimization objective or convergence.
The paper provides no convergence analysis or ablation showing the scaling rule is stable across training.

how reviews work

0 comments

Cite this review

Pith. "Pith review of BEAT: Balanced Frequency Adaptive Tuning for Long-Term Time-Series Forecasting." pith.science (2026). https://pith.science/paper/WSMDS23C

@misc{pith2026250119065,
  author       = {Pith},
  title        = {Pith review of: BEAT: Balanced Frequency Adaptive Tuning for Long-Term Time-Series Forecasting},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WSMDS23C}},
  note         = {Machine review of arXiv:2501.19065}
}

read the original abstract

Time-series forecasting is crucial for numerous real-world applications including weather prediction and financial market modeling. While temporal-domain methods remain prevalent, frequency-domain approaches can effectively capture multi-scale periodic patterns, reduce sequence dependencies, and naturally denoise signals. However, existing approaches typically train model components for all frequencies under a unified training objective, often leading to mismatched learning speeds: high-frequency components converge faster and risk overfitting, while low-frequency components underfit due to insufficient training time. To deal with this challenge, we propose BEAT (Balanced frEquency Adaptive Tuning), a novel framework that dynamically monitors the training status for each frequency and adaptively adjusts their gradient updates. By recognizing convergence, overfitting, or underfitting for each frequency, BEAT dynamically reallocates learning priorities, moderating gradients for rapid learners and increasing those for slower ones, alleviating the tension between competing objectives across frequencies and synchronizing the overall learning process. Extensive experiments on seven real-world datasets demonstrate that BEAT consistently outperforms state-of-the-art approaches.

Figures

Figures reproduced from arXiv: 2501.19065 by the authors.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

47 extracted references · 34 canonical work pages

[1]

Z.; and Koltun, V

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

arXiv 2018
[2]

Brigham, E. 1988. The fast Fourier transform and its applications

work page 1988
[3]

Das, A.; Kong, W.; Leach, A.; Mathur, S.; Sen, R.; and Yu, R. 2023. Long-term forecasting with TiDE : Time-series dense encoder. In arXiv preprint arXiv:2304.08424

arXiv 2023
[4]

Daubechies, I. 1992. Ten lectures on wavelets. Society for industrial and applied mathematics

work page 1992
[5]

Ekambaram, V.; Jati, A.; Nguyen, N.; Sinthong, P.; and Kalagnanam, J. 2023. TSMixer : Lightweight mlp-mixer model for multivariate time series forecasting. In Proceedings of the ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 459--469

work page 2023
[6]

Eldele, E.; Ragab, M.; Chen, Z.; Wu, M.; and Li, X. 2024. Tslanet: Rethinking transformers for time series representation learning. arXiv preprint arXiv:2404.08472

arXiv 2024
[7]

Gasparin, A.; Lukovic, S.; and Alippi, C. 2022. Deep learning for time series forecasting: The electric load case. CAAI Transactions on Intelligence Technology, 7(1): 1--25

work page 2022
[8]

Gu, J.; Wang, Z.; Kuen, J.; Ma, L.; Shahroudy, A.; Shuai, B.; Liu, T.; Wang, X.; Wang, G.; Cai, J.; et al. 2018. Recent advances in convolutional neural networks. Pattern recognition, 77: 354--377

work page 2018

Show all 47 references

[9]

He, K.; Zhang, X.; Ren, S.; and Sun, J. 2016. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, 770--778

2016
[10]

Jia, F.; Wang, K.; Zheng, Y.; Cao, D.; and Liu, Y. 2024. GPT4MTS : Prompt-based Large Language Model for Multimodal Time-series Forecasting. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, 23343--23351

2024
[11]

Jiang, X.; Jin, S.; Lu, L.; Zhang, X.; and Lu, S. 2024 a . Weakly Supervised Monocular 3 D Detection with a Single-View Image. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 10508--10518

2024
[12]

Jiang, X.; Jin, S.; Zhang, X.; Shao, L.; and Lu, S. 2024 b . MonoMAE: Enhancing Monocular 3 D Detection through Depth-Aware Masked Autoencoders. In Advances in Neural Information Processing Systems

2024
[13]

Jin, K.; Wi, J.; Lee, E.; Kang, S.; Kim, S.; and Kim, Y. 2021. TrafficBERT : Pre-trained model with large-scale data for long-range traffic flow forecasting. Expert Systems with Applications, 186: 115738

2021
[14]

Y.; Shi, X.; Chen, P.-Y.; Liang, Y.; Li, Y.-F.; Pan, S.; and Wen, Q

Jin, M.; Wang, S.; Ma, L.; Chu, Z.; Zhang, J. Y.; Shi, X.; Chen, P.-Y.; Liang, Y.; Li, Y.-F.; Pan, S.; and Wen, Q. 2024. Time-LLM : Time Series Forecasting by Reprogramming Large Language Models. In International Conference on Learning Representations

2024
[15]

Kim, T.; Kim, J.; Tae, Y.; Park, C.; Choi, J.-H.; and Choo, J. 2021. Reversible instance normalization for accurate time-series forecasting against distribution shift. In International Conference on Learning Representations

2021
[16]

Lai, G.; Chang, W.-C.; Yang, Y.; and Liu, H. 2018. Modeling long-and short-term temporal patterns with deep neural networks. In The 41st international ACM SIGIR conference on research & development in information retrieval, 95--104

2018
[17]

Li, S.; Jin, X.; Xuan, Y.; Zhou, X.; Chen, W.; Wang, Y.-X.; and Yan, X. 2019. Enhancing the locality and breaking the memory bottleneck of transformer on time series forecasting. Advances in Neural Information Processing Systems, 32

2019
[18]

Li, Z.; Ye, W.; Jiang, T.; and Huang, T. 2022. 2 D Amodal Instance Segmentation Guided by 3 D Shape Prior. In Proceedings of the IEEE/CVF European Conference on Computer Vision, 165--181

2022
[19]

Li, Z.; Ye, W.; Terven, J.; Bennett, Z.; Zheng, Y.; Jiang, T.; and Huang, T. 2023. MUVA : A New Large-Scale Benchmark for Multi-view Amodal Instance Segmentation in the Shopping Scenario. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 23504--23513

2023
[20]

Liu, M.; Zeng, A.; Chen, M.; Xu, Z.; Lai, Q.; Ma, L.; and Xu, Q. 2022. SCINet : Time series modeling and forecasting with sample convolution and interaction. Advances in Neural Information Processing Systems, 35: 5816--5828

2022
[21]

X.; and Dustdar, S

Liu, S.; Yu, H.; Liao, C.; Li, J.; Lin, W.; Liu, A. X.; and Dustdar, S. 2021. Pyraformer: Low-complexity pyramidal attention for long-range time series modeling and forecasting. In International Conference on Learning Representations

2021
[22]

Liu, Y.; Hu, T.; Zhang, H.; Wu, H.; Wang, S.; Ma, L.; and Long, M. 2024. iTransformer: Inverted transformers are effective for time series forecasting. In International Conference on Learning Representations

2024
[23]

Luo, D.; and Wang, X. 2024. ModernTCN : A modern pure convolution structure for general time series analysis. In International Conference on Learning Representations

2024
[24]

Mallat, S. G. 1989. A theory for multiresolution signal decomposition: the wavelet representation. IEEE transactions on pattern analysis and machine intelligence, 11(7): 674--693

1989
[25]

Murad, M. M. N.; Aktukmak, M.; and Yilmaz, Y. 2025. WPMixer : Efficient Multi-Resolution Mixing for Long-Term Time Series Forecasting. In Proceedings of the AAAI Conference on Artificial Intelligence

2025
[26]

H.; Sinthong, P.; and Kalagnanam, J

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

2023
[27]

N.; Carpov, D.; Chapados, N.; and Bengio, Y

Oreshkin, B. N.; Carpov, D.; Chapados, N.; and Bengio, Y. 2019. N-BEATS : Neural basis expansion analysis for interpretable time series forecasting. arXiv preprint arXiv:1905.10437

2019 arXiv
[28]

Paszke, A.; Gross, S.; Massa, F.; Lerer, A.; Bradbury, J.; Chanan, G.; Killeen, T.; Lin, Z.; Gimelshein, N.; Antiga, L.; et al. 2019. Pytorch: An imperative style, high-performance deep learning library. Advances in neural information processing systems, 32

2019
[29]

Peng, X.; Wei, Y.; Deng, A.; Wang, D.; and Hu, D. 2022. Balanced multimodal learning via on-the-fly gradient modulation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 8238--8247

2022
[30]

Tang, Y.; Song, Z.; Zhu, Y.; Yuan, H.; Hou, M.; Ji, J.; Tang, C.; and Li, J. 2022. A survey on machine learning models for financial time series forecasting. Neurocomputing, 512: 363--380

2022
[31]

Vaswani, A. 2017. Attention is all you need. Advances in Neural Information Processing Systems

2017
[32]

Wang, H.; Peng, J.; Huang, F.; Wang, J.; Chen, J.; and Xiao, Y. 2023. MICN : Multi-scale local and global context modeling for long-term series forecasting. In International Conference on Learning Representations

2023
[33]

Wang, J.; Wang, Z.; Li, J.; and Wu, J. 2018. Multilevel wavelet decomposition network for interpretable time series analysis. In Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, 2437--2446

2018
[34]

Y.; and ZHOU, J

Wang, S.; Wu, H.; Shi, X.; Hu, T.; Luo, H.; Ma, L.; Zhang, J. Y.; and ZHOU, J. 2024 a . TimeMixer: Decomposable Multiscale Mixing for Time Series Forecasting. In International Conference on Learning Representations

2024
[35]

Wang, Y.; Wu, H.; Dong, J.; Liu, Y.; Long, M.; and Wang, J. 2024 b . Deep Time Series Models: A Comprehensive Survey and Benchmark

2024
[36]

Wang, Z.; Pei, C.; Ma, M.; Wang, X.; Li, Z.; Pei, D.; Rajmohan, S.; Zhang, D.; Lin, Q.; Zhang, H.; et al. 2024 c . Revisiting VAE for Unsupervised Time Series Anomaly Detection: A Frequency Perspective. In Proceedings of the ACM on Web Conference 2024, 3096--3105

2024
[37]

Wu, H.; Hu, T.; Liu, Y.; Zhou, H.; Wang, J.; and Long, M. 2023. TimesNet : Temporal 2D-Variation Modeling for General Time Series Analysis. In International Conference on Learning Representations

2023
[38]

Wu, H.; Xu, J.; Wang, J.; and Long, M. 2021. Autoformer: Decomposition transformers with auto-correlation for long-term series forecasting. In Advances in Neural Information Processing Systems, volume 34, 22419--22430

2021
[39]

Xu, Z.; Zeng, A.; and Xu, Q. 2023. FITS: Modeling time series with 10 k parameters. arXiv preprint arXiv:2307.03756

2023 arXiv
[40]

Yi, K.; Zhang, Q.; Fan, W.; Wang, S.; Wang, P.; He, H.; An, N.; Lian, D.; Cao, L.; and Niu, Z. 2024. Frequency-domain MLPs are more effective learners in time series forecasting. Advances in Neural Information Processing Systems, 36

2024
[41]

Yi, K.; Zhang, Q.; Hu, L.; He, H.; An, N.; Cao, L.; and Niu, Z. 2022. Edge-Varying Fourier Graph Networks for Multivariate Time Series Forecasting. arXiv preprint arXiv:2210.03093

2022 arXiv
[42]

Zeng, A.; Chen, M.; Zhang, L.; and Xu, Q. 2023. Are transformers effective for time series forecasting? In Proceedings of the AAAI Conference on Artificial Intelligence, volume 37, 11121--11128

2023
[43]

Zhang, G.; Yang, D.; Galanis, G.; and Androulakis, E. 2022. Solar forecasting with hourly updated numerical weather prediction. Renewable and Sustainable Energy Reviews, 154: 111768

2022
[44]

Zhang, Y.; and Yan, J. 2023. Crossformer: Transformer utilizing cross-dimension dependency for multivariate time series forecasting. In International Conference on Learning Representations

2023
[45]

Zhou, H.; Zhang, S.; Peng, J.; Zhang, S.; Li, J.; Xiong, H.; and Zhang, W. 2021. Informer: Beyond efficient transformer for long sequence time-series forecasting. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 35, 11106--11115

2021
[46]

Zhou, T.; Ma, Z.; Wen, Q.; Sun, L.; Yao, T.; Yin, W.; Jin, R.; et al. 2022 a . Film: Frequency improved legendre memory model for long-term time series forecasting. Advances in neural information processing systems, 35: 12677--12690

2022
[47]

Zhou, T.; Ma, Z.; Wen, Q.; Wang, X.; Sun, L.; and Jin, R. 2022 b . FEDFormer : Frequency enhanced decomposed transformer for long-term series forecasting. In International Conference on Machine Learning, 27268--27286

2022

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.

[1] [1]

Z.; and Koltun, V

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

arXiv 2018

[2] [2]

Brigham, E. 1988. The fast Fourier transform and its applications

work page 1988

[3] [3]

Das, A.; Kong, W.; Leach, A.; Mathur, S.; Sen, R.; and Yu, R. 2023. Long-term forecasting with TiDE : Time-series dense encoder. In arXiv preprint arXiv:2304.08424

arXiv 2023

[4] [4]

Daubechies, I. 1992. Ten lectures on wavelets. Society for industrial and applied mathematics

work page 1992

[5] [5]

Ekambaram, V.; Jati, A.; Nguyen, N.; Sinthong, P.; and Kalagnanam, J. 2023. TSMixer : Lightweight mlp-mixer model for multivariate time series forecasting. In Proceedings of the ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 459--469

work page 2023

[6] [6]

Eldele, E.; Ragab, M.; Chen, Z.; Wu, M.; and Li, X. 2024. Tslanet: Rethinking transformers for time series representation learning. arXiv preprint arXiv:2404.08472

arXiv 2024

[7] [7]

Gasparin, A.; Lukovic, S.; and Alippi, C. 2022. Deep learning for time series forecasting: The electric load case. CAAI Transactions on Intelligence Technology, 7(1): 1--25

work page 2022

[8] [8]

Gu, J.; Wang, Z.; Kuen, J.; Ma, L.; Shahroudy, A.; Shuai, B.; Liu, T.; Wang, X.; Wang, G.; Cai, J.; et al. 2018. Recent advances in convolutional neural networks. Pattern recognition, 77: 354--377

work page 2018