pith. sign in

arxiv: 2606.23425 · v1 · pith:Z4DEXDS4new · submitted 2026-06-22 · 💻 cs.LG · cs.CE· cs.SC

Interpretable Kolmogorov-Arnold Network with Feature-Isolated Temporal Attention Mechanism for Electricity Load Forecasting

Jinhao Li , Hao Wang This is my paper

Pith reviewed 2026-06-26 08:33 UTC · model grok-4.3

classification 💻 cs.LG cs.CEcs.SC
keywords electricity load forecastingKolmogorov-Arnold Networkinterpretabilitytemporal attentionhuman mobilitytime series forecasting
0
0 comments X

The pith

A hybrid attention-KAN model forecasts electricity load competitively while revealing interpretable mobility effects.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper presents LoadKAN, which uses a feature-isolated temporal attention mechanism to process each input feature's temporal dynamics separately before passing distilled representations to a Kolmogorov-Arnold Network. This design allows the KAN's learnable activation functions to expose non-linear relationships between mobility patterns and load. On three U.S. electricity market datasets, LoadKAN performs on par with extensively tuned black-box deep learning models. The interpretability feature supports detailed sensitivity analysis of how different mobility patterns affect demand in each market. Such transparency addresses a key limitation of standard neural forecasting approaches.

Core claim

LoadKAN is a novel hybrid framework that combines a feature-isolated temporal attention mechanism with a KAN module; the attention stage extracts temporal dynamics from each input feature independently to supply the KAN with representations for interpretable predictions, and evaluations show it competes with state-of-the-art black-box models while enabling granular analysis of non-linear mobility-load relationships through KAN-learned activation functions.

What carries the argument

Feature-isolated temporal attention mechanism that independently extracts temporal dynamics from each input feature to feed the KAN module.

If this is right

  • Remains competitive with state-of-the-art black-box deep learning models on three U.S. electricity market datasets.
  • Enables granular analysis of learned non-linear relationships between six mobility patterns and electricity load.
  • Supports quantitative sensitivity analyses revealing complex and market-specific dependencies on mobility features.
  • Generates insights often hidden in black-box neural forecasting models.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Similar hybrid designs could improve interpretability in other forecasting tasks involving multiple input streams.
  • The market-specific mobility dependencies indicate that general models may miss local patterns in load prediction.
  • Isolating features in attention might help reduce spurious correlations in time-series models more broadly.

Load-bearing premise

The feature-isolated temporal attention can extract temporal dynamics from each feature independently to give the KAN representations that produce reliable insights into mobility-load relationships.

What would settle it

A test showing that the activation functions learned by the KAN do not correspond to verifiable patterns in the mobility data or that LoadKAN accuracy falls below that of the compared black-box models on the evaluation datasets.

Figures

Figures reproduced from arXiv: 2606.23425 by Hao Wang, Jinhao Li.

Figure 1
Figure 1. Figure 1: Load profiles of NYISO, CAISO, and ERCOT. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Mobility features of CAISO, NYISO, and ERCOT. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: The LoadKAN framework. stages: first, a non-entangled temporal feature processing stage captures the dynamics of each input feature sequence indepen￾dently; second, a KAN module models the relationships be￾tween these processed feature representations and the final load forecast. The following section provides a rigorous mathemati￾cal formulation of each module within LoadKAN. For a specific forecast initi… view at source ↗
Figure 4
Figure 4. Figure 4: Performance improvements after integrating mobility features. [PITH_FULL_IMAGE:figures/full_fig_p010_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Visualizations of learned activation functions in the NYISO market. [PITH_FULL_IMAGE:figures/full_fig_p012_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Visualizations of learned activation functions in the CAISO market. [PITH_FULL_IMAGE:figures/full_fig_p012_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Visualizations of learned activation functions in the ERCOT market. [PITH_FULL_IMAGE:figures/full_fig_p012_7.png] view at source ↗
Figure 10
Figure 10. Figure 10: Sensitivity of learned activation functions in the ERCOT market. [PITH_FULL_IMAGE:figures/full_fig_p013_10.png] view at source ↗
read the original abstract

Accurate electricity load forecasting is a crucial prerequisite for stable power system operations. While prevalent deep learning models present competitive performance, they often operate as black boxes and lack interpretability. While the Kolmogorov-Arnold network (KAN) has emerged as a promising alternative because of its learnable activation function design, its direct application to time-series forecasting faces challenges in modeling complex temporal data patterns. Also, simple integration into existing architectures, such as serving as replacement of neural modules, cannot fully leverage KAN's interpretability strengths. To address these gaps, this study develops LoadKAN, a novel hybrid and interpretable framework for load forecasting that synergistically combines a specifically-designed feature-isolated temporal attention mechanism with a KAN module. The attention stage aims to extract temporal dynamics from each input feature independently, such as historical load and human mobility, providing distilled feature representations to the KAN module for interpretable predictions. When evaluated on datasets from three representative U.S. electricity markets, our LoadKAN remains highly competitive when compared to extensively-tuned, state-of-the-art, black-box deep learning benchmarks. More importantly, LoadKAN's interpretability enables a granular analysis of the learned non-linear relationships between six distinct mobility patterns and electricity load. Through KAN-learned activation functions, our quantitative sensitivity analyses on mobility features reveal complex and market-specific dependencies. These findings further demonstrate the ability of our LoadKAN to generate insights often obscured by opaque black-box neural forecasting models.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 0 minor

Summary. The paper proposes LoadKAN, a hybrid architecture that pairs a feature-isolated temporal attention stage (intended to distill independent temporal dynamics from each input feature such as historical load and six mobility patterns) with a Kolmogorov-Arnold Network (KAN) module. It claims that the resulting model remains competitive with extensively tuned black-box deep-learning baselines on three U.S. electricity-market datasets while additionally supplying interpretable univariate activation functions that reveal market-specific non-linear mobility-load relationships.

Significance. If the performance comparisons are statistically sound and the KAN activations can be shown to reflect genuine rather than artifactual structure, the work would offer a concrete route to interpretable load forecasting that exploits mobility data. The interpretability component, however, rests on an unvalidated assumption that the attention stage supplies the KAN with representations free of residual temporal leakage or spurious correlations; without supporting ablations or external checks this contribution remains provisional.

major comments (2)
  1. [Abstract] Abstract: the claim of 'highly competitive' performance against 'extensively-tuned, state-of-the-art, black-box deep learning benchmarks' is asserted without any reported metrics, baseline specifications, statistical tests, or error analysis, rendering the central empirical claim impossible to evaluate from the provided text.
  2. [Abstract] Abstract (interpretability paragraph): the assertion that 'KAN-learned activation functions' and 'quantitative sensitivity analyses on mobility features reveal complex and market-specific dependencies' is load-bearing for the interpretability contribution, yet no ablation of the feature-isolated attention, consistency check across random seeds, or external validation against known mobility-electricity elasticities is described; the design alone does not guarantee that the univariate functions are non-spurious.

Simulated Author's Rebuttal

2 responses · 1 unresolved

We thank the referee for the constructive feedback. We agree that the abstract would benefit from greater specificity to support its claims and will revise it accordingly. We address each major comment below.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the claim of 'highly competitive' performance against 'extensively-tuned, state-of-the-art, black-box deep learning benchmarks' is asserted without any reported metrics, baseline specifications, statistical tests, or error analysis, rendering the central empirical claim impossible to evaluate from the provided text.

    Authors: We acknowledge that the abstract does not include quantitative details. The full manuscript (Experiments section) reports MAE, RMSE, and MAPE on the three U.S. market datasets, with comparisons to extensively tuned LSTM, GRU, Transformer, and other baselines, including hyperparameter search procedures and error analysis. Statistical significance across random seeds is assessed via paired tests. In revision we will condense these results into the abstract (e.g., 'LoadKAN achieves X% lower MAE than the best baseline on average') while retaining the high-level claim. revision: yes

  2. Referee: [Abstract] Abstract (interpretability paragraph): the assertion that 'KAN-learned activation functions' and 'quantitative sensitivity analyses on mobility features reveal complex and market-specific dependencies' is load-bearing for the interpretability contribution, yet no ablation of the feature-isolated attention, consistency check across random seeds, or external validation against known mobility-electricity elasticities is described; the design alone does not guarantee that the univariate functions are non-spurious.

    Authors: The manuscript presents the KAN activation functions and sensitivity plots in the Results section, showing market-specific non-linear patterns. We agree the abstract does not reference supporting checks. In revision we will add a sentence pointing to the relevant figures and note that the feature-isolated attention processes each input stream independently before the KAN stage. We will also include a short consistency analysis across seeds. External validation against published mobility-electricity elasticities is not performed in the current work and would require additional domain datasets; we will therefore qualify the claim to reflect the evidence actually provided. revision: partial

standing simulated objections not resolved
  • External validation of the learned mobility-load relationships against independent econometric elasticities, which is outside the scope of the present datasets and would require new data collection.

Circularity Check

0 steps flagged

No circularity; empirical evaluation independent of self-defined quantities

full rationale

The paper describes a hybrid architecture (feature-isolated temporal attention feeding a KAN module) and reports empirical performance on three external U.S. electricity market datasets against black-box benchmarks. No derivation chain, equations, or fitted-parameter predictions are presented that reduce to the model's own inputs by construction. Interpretability claims rest on post-hoc inspection of learned KAN activations rather than any self-referential prediction or uniqueness theorem imported via self-citation. The central claims therefore remain externally falsifiable and do not collapse into the architecture's fitted values.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Only the abstract is available; no equations, parameter counts, or modeling assumptions are stated, so the ledger remains empty.

pith-pipeline@v0.9.1-grok · 5797 in / 1026 out tokens · 24443 ms · 2026-06-26T08:33:48.898216+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

68 extracted references · 43 canonical work pages

  1. [1]

    El Maghraoui, H

    A. El Maghraoui, H. El Hadraoui, Y . Ledmaoui, N. El Bazi, N. Guennouni, A. Chebak, Revolutionizing smart 19 grid-ready management systems: A holistic frame- work for optimal grid reliability, Sustainable Energy, Grids and Networks 39 (2024) 101452.doi:https: //doi.org/10.1016/j.segan.2024.101452. URLhttps://www.sciencedirect.com/science/ article/pii/S235...

    work page doi:10.1016/j.segan.2024.101452 2024

  2. [3]

    L. Yang, R. Ren, X. Gu, L. Sun, Interactive general- ized additive model and its applications in electric load forecasting, in: Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD ’23, Association for Computing Machinery, New York, NY , USA, 2023, p. 5393–5403.doi:10.1145/ 3580305.3599848. URLhttps://doi.org/10.1145/3...

    work page doi:10.1145/3580305.3599848 2023

  3. [4]

    Marshman, M

    D. Marshman, M. Brear, B. Ring, Impact of unit com- mitment and rocof constraints on revenue sufficiency in decarbonising wholesale electricity markets, En- ergy Economics 106 (2022) 105763.doi:https: //doi.org/10.1016/j.eneco.2021.105763. URLhttps://www.sciencedirect.com/science/ article/pii/S0140988321006046

    work page doi:10.1016/j.eneco.2021.105763 2022

  4. [5]

    Davis, M

    D. Davis, M. J. Brear, Impact of short-term wind forecast accuracy on the performance of decar- bonising wholesale electricity markets, Energy Economics 130 (2024) 107304.doi:https: //doi.org/10.1016/j.eneco.2024.107304. URLhttps://www.sciencedirect.com/science/ article/pii/S0140988324000124

    work page doi:10.1016/j.eneco.2024.107304 2024

  5. [6]

    Next Move Strategy Consulting, Power generation market – global opportunity analysis and industry forecast, 2024–2030,https://www.nextmsc.com/report/ power-generation-market, accessed: 2025-05-09 (2024)

    2024

  6. [7]

    Licence: CC BY 4.0

    International Energy Agency, Electricity 2024,https: //www.iea.org/reports/electricity-2024, iEA, Paris. Licence: CC BY 4.0. Accessed: 2025-05-09 (2024)

    2024

  7. [8]

    Rubasinghe, T

    O. Rubasinghe, T. Zhang, X. Zhang, S. S. Choi, T. K. Chau, Y . Chow, T. Fernando, H. H.-C. Iu, Highly accurate peak and valley prediction short-term net load forecasting approach based on decompo- sition for power systems with high pv penetration, Applied Energy 333 (2023) 120641.doi:https: //doi.org/10.1016/j.apenergy.2023.120641. URLhttps://www.scienced...

    work page doi:10.1016/j.apenergy.2023.120641 2023

  8. [9]

    H. Min, B. Noh, Solarnexus: A deep learning framework for adaptive photovoltaic power gen- eration forecasting and scalable management, Ap- plied Energy 391 (2025) 125848.doi:https: //doi.org/10.1016/j.apenergy.2025.125848. URLhttps://www.sciencedirect.com/science/ article/pii/S0306261925005781

    work page doi:10.1016/j.apenergy.2025.125848 2025

  9. [10]

    X. Wang, H. Wang, S. Li, H. Jin, A reinforcement learning-based online learning strategy for real-time short-term load forecasting, Energy 305 (2024) 132344

    2024

  10. [11]

    Kuster, Y

    C. Kuster, Y . Rezgui, M. Mourshed, Electrical load fore- casting models: A critical systematic review, Sustainable Cities and Society 35 (2017) 257–270.doi:https: //doi.org/10.1016/j.scs.2017.08.009. URLhttps://www.sciencedirect.com/science/ article/pii/S2210670717305899

    work page doi:10.1016/j.scs.2017.08.009 2017

  11. [12]

    M. Xia, H. Shao, X. Ma, C. W. de Silva, A stacked gru- rnn-based approach for predicting renewable energy and electricity load for smart grid operation, IEEE Transac- tions on Industrial Informatics 17 (10) (2021) 7050–7059. doi:10.1109/TII.2021.3056867

    work page doi:10.1109/tii.2021.3056867 2021

  12. [13]

    C. Wang, H. Zhao, Y . Liu, G. Fan, Minute-level ultra- short-term power load forecasting based on time series data features, Applied Energy 372 (2024) 123801. doi:https://doi.org/10.1016/j.apenergy. 2024.123801. URLhttps://www.sciencedirect.com/science/ article/pii/S030626192401184X

    work page doi:10.1016/j.apenergy 2024

  13. [14]

    J. Li, H. Wang, A multi-view multi-timescale hypergraph- empowered spatiotemporal framework for ev charging forecasting, IEEE Transactions on Smart Grid 17 (2) (2026) 1430–1443.doi:10.1109/TSG.2025.3633777

    work page doi:10.1109/tsg.2025.3633777 2026

  14. [15]

    Mansoor, M

    H. Mansoor, M. S. Gull, H. Rauf, I. ul Hasan Shaikh, M. Khalid, N. Arshad, Graph convolutional networks based short-term load forecasting: Leveraging spatial information for improved accuracy, Electric Power Systems Research 230 (2024) 110263.doi:https: //doi.org/10.1016/j.epsr.2024.110263. URLhttps://www.sciencedirect.com/science/ article/pii/S0378779624001512

    work page doi:10.1016/j.epsr.2024.110263 2024

  15. [16]

    Giamarelos, E

    N. Giamarelos, E. N. Zois, Boosting short term electric load forecasting of high & medium voltage substations with visibility graphs and graph neural networks, Sustain- able Energy, Grids and Networks 38 (2024) 101304.doi: https://doi.org/10.1016/j.segan.2024.101304. URLhttps://www.sciencedirect.com/science/ article/pii/S235246772400033X

    work page doi:10.1016/j.segan.2024.101304 2024

  16. [17]

    Z. Peng, X. Yang, Short- and medium-term power load forecasting model based on a hybrid attention mechanism in the time and frequency domains, Expert Systems with Applications 278 (2025) 127329.doi:https: //doi.org/10.1016/j.eswa.2025.127329. 20 URLhttps://www.sciencedirect.com/science/ article/pii/S0957417425009510

    work page doi:10.1016/j.eswa.2025.127329 2025

  17. [18]

    Jiang, H

    B. Jiang, H. Yang, Y . Wang, Y . Liu, H. Geng, H. Zeng, J. Ding, Dynamic temporal dependency model for multi- ple steps ahead short-term load forecasting of power sys- tem, IEEE Transactions on Industry Applications 60 (4) (2024) 5244–5254.doi:10.1109/TIA.2024.3375802

    work page doi:10.1109/tia.2024.3375802 2024

  18. [19]

    Y . Chen, W. Yang, B. Zhang, Using mobility for electri- cal load forecasting during the covid-19 pandemic (2020). arXiv:2006.08826. URLhttps://arxiv.org/abs/2006.08826

    arXiv 2020

  19. [20]

    J. Liu, Z. Zhang, X. Fan, Y . Zhang, J. Wang, K. Zhou, S. Liang, X. Yu, W. Zhang, Power system load forecasting using mobility optimiza- tion and multi-task learning in covid-19, Ap- plied Energy 310 (2022) 118303.doi:https: //doi.org/10.1016/j.apenergy.2021.118303. URLhttps://www.sciencedirect.com/science/ article/pii/S0306261921015622

    work page doi:10.1016/j.apenergy.2021.118303 2022

  20. [21]

    Zarbakhsh, M

    N. Zarbakhsh, M. S. Misaghian, G. Mcardle, Human mobility-based features to analyse the impact of covid- 19 on power system operation of ireland, IEEE Open Access Journal of Power and Energy 9 (2022) 213–225. doi:10.1109/OAJPE.2022.3155960

    work page doi:10.1109/oajpe.2022.3155960 2022

  21. [22]

    Cordova, L

    J. Cordova, L. M. Konila Sriram, A. Kocatepe, Y . Zhou, E. E. Ozguven, R. Arghandeh, Combined electricity and traffic short-term load forecasting using bundled causality engine, IEEE Transactions on Intelligent Transportation Systems 20 (9) (2019) 3448–3458.doi:10.1109/TITS. 2018.2876871

    work page doi:10.1109/tits 2019

  22. [23]

    Zhang, J

    Z. Zhang, J. Liu, S. Pang, M. Shi, H. H. Goh, Y . Zhang, D. Zhang, General short-term load forecasting based on multi-task temporal convolutional network in covid-19, International Journal of Electrical Power & Energy Systems 147 (2023) 108811.doi:https: //doi.org/10.1016/j.ijepes.2022.108811. URLhttps://www.sciencedirect.com/science/ article/pii/S0142061...

    work page doi:10.1016/j.ijepes.2022.108811 2023

  23. [24]

    Prabowo, K

    A. Prabowo, K. Chen, H. Xue, S. Sethuvenkatraman, F. D. Salim, Navigating out-of-distribution electricity load fore- casting during covid-19: Benchmarking energy load fore- casting models without and with continual learning, in: Proceedings of the 10th ACM International Conference on Systems for Energy-Efficient Buildings, Cities, and Transportation, Buil...

    work page doi:10.1145/3600100.3623726 2023

  24. [25]

    D. Obst, J. de Vilmarest, Y . Goude, Adaptive methods for short-term electricity load forecasting during covid-19 lockdown in france, IEEE Transactions on Power Systems 36 (5) (2021) 4754–4763.doi:10.1109/TPWRS.2021. 3067551

    work page doi:10.1109/tpwrs.2021 2021

  25. [26]

    Z. Liu, Y . Wang, S. Vaidya, F. Ruehle, J. Halver- son, M. Soljacic, T. Y . Hou, M. Tegmark, KAN: Kolmogorov–arnold networks, in: The Thirteenth Inter- national Conference on Learning Representations, 2025. URLhttps://openreview.net/forum?id= Ozo7qJ5vZi

    2025

  26. [27]

    L. Baur, K. Ditschuneit, M. Schambach, C. Kaymakci, T. Wollmann, A. Sauer, Explainability and interpretability in electric load forecasting using machine learning tech- niques – a review, Energy and AI 16 (2024) 100358.doi: https://doi.org/10.1016/j.egyai.2024.100358. URLhttps://www.sciencedirect.com/science/ article/pii/S2666546824000247

    work page doi:10.1016/j.egyai.2024.100358 2024

  27. [28]

    Why Should I Trust You?

    M. T. Ribeiro, S. Singh, 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, KDD ’16, Association for Computing Machinery, New York, NY , USA, 2016, p. 1135–1144.doi:10.1145/ 2939672.2939778. URLhttps://doi.org/10.1145/...

    work page doi:10.1145/2939672.2939778 2016

  28. [29]

    S. M. Lundberg, S.-I. Lee, A unified approach to inter- preting model predictions, in: I. Guyon, U. V . Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, R. Garnett (Eds.), Advances in Neural Information Processing Systems 30, Curran Associates, Inc., 2017, pp. 4765–4774. URLhttp://papers.nips.cc/paper/ 7062-a-unified-approach-to-interpreting-mo...

    2017

  29. [30]

    Slack, A

    D. Slack, A. Hilgard, S. Singh, H. Lakkaraju, Reliable post hoc explanations: Modeling uncertainty in explain- ability, in: M. Ranzato, A. Beygelzimer, Y . Dauphin, P. Liang, J. W. Vaughan (Eds.), Advances in Neural Infor- mation Processing Systems, V ol. 34, Curran Associates, Inc., 2021, pp. 9391–9404. URLhttps://proceedings.neurips. cc/paper_files/pape...

    2021

  30. [31]

    Waheed, Q

    W. Waheed, Q. Xu, Data-driven short term load fore- casting with deep neural networks: Unlocking insights for sustainable energy management, Electric Power Systems Research 232 (2024) 110376.doi:https: //doi.org/10.1016/j.epsr.2024.110376. URLhttps://www.sciencedirect.com/science/ article/pii/S0378779624002645

    work page doi:10.1016/j.epsr.2024.110376 2024

  31. [32]

    Long short-term memory.Neural Computation, 9(8):1735–1780, 1997

    S. Hochreiter, J. Schmidhuber, Long short-term memory, Neural Comput. 9 (8) (1997) 1735–1780. doi:10.1162/neco.1997.9.8.1735. URLhttps://doi.org/10.1162/neco.1997.9.8. 1735 21

    work page doi:10.1162/neco.1997.9.8.1735 1997

  32. [33]

    Chung, C

    J. Chung, C. Gulcehre, K. Cho, Y . Bengio, Empirical eval- uation of gated recurrent neural networks on sequence modeling, cite arxiv:1412.3555Comment: Presented in NIPS 2014 Deep Learning and Representation Learning Workshop (2014). URLhttp://arxiv.org/abs/1412.3555

    Pith/arXiv arXiv 2014

  33. [34]

    C. Lea, M. D. Flynn, R. Vidal, A. Reiter, G. D. Hager, Temporal convolutional networks for action segmentation and detection, in: 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017, pp. 1003– 1012.doi:10.1109/CVPR.2017.113

    work page doi:10.1109/cvpr.2017.113 2017

  34. [35]

    Eskandari, M

    H. Eskandari, M. Imani, M. P. Moghaddam, Convo- lutional and recurrent neural network based model for short-term load forecasting, Electric Power Sys- tems Research 195 (2021) 107173.doi:https: //doi.org/10.1016/j.epsr.2021.107173. URLhttps://www.sciencedirect.com/science/ article/pii/S0378779621001541

    work page doi:10.1016/j.epsr.2021.107173 2021

  35. [36]

    Y . Chen, C. Lin, Y . Zhang, J. Liu, D. Yu, Day-ahead load forecast based on conv2d-gru_sc aimed to adapt to steep changes in load, Energy 302 (2024) 131814.doi:https: //doi.org/10.1016/j.energy.2024.131814. URLhttps://www.sciencedirect.com/science/ article/pii/S0360544224015871

    work page doi:10.1016/j.energy.2024.131814 2024

  36. [37]

    Saeed, A

    F. Saeed, A. Rehman, H. A. Shah, M. Diyan, J. Chen, J.-M. Kang, Smartformer: Graph-based transformer model for energy load forecasting, Sustainable Energy Technologies and Assessments 73 (2025) 104133.doi: https://doi.org/10.1016/j.seta.2024.104133. URLhttps://www.sciencedirect.com/science/ article/pii/S2213138824005290

    work page doi:10.1016/j.seta.2024.104133 2025

  37. [38]

    Zhang, Y

    C. Zhang, Y . Yu, T. Zhang, K. Song, Y . Wang, S. Gao, Short-term load forecasting based on graph convolu- tion and dendritic deep learning, IEEE Transactions on Network Science and Engineering (2025) 1–13doi:10. 1109/TNSE.2025.3558193

    arXiv 2025

  38. [39]

    Z. Su, G. Zheng, M. Hu, L. Kong, G. Wang, Short-term load forecasting of regional integrated energy system based on spatio-temporal convolu- tional graph neural network, Electric Power Sys- tems Research 232 (2024) 110427.doi:https: //doi.org/10.1016/j.epsr.2024.110427. URLhttps://www.sciencedirect.com/science/ article/pii/S0378779624003158

    work page doi:10.1016/j.epsr.2024.110427 2024

  39. [40]

    Z. Tian, W. Liu, W. Jiang, C. Wu, Cnns- transformer based day-ahead probabilistic load forecasting for weekends with limited data avail- ability, Energy 293 (2024) 130666.doi:https: //doi.org/10.1016/j.energy.2024.130666. URLhttps://www.sciencedirect.com/science/ article/pii/S0360544224004389

    work page doi:10.1016/j.energy.2024.130666 2024

  40. [41]

    H. Xu, F. Hu, X. Liang, G. Zhao, M. Abugunmi, A frame- work for electricity load forecasting based on attention mechanism time series depthwise separable convolutional neural network, Energy 299 (2024) 131258.doi:https: //doi.org/10.1016/j.energy.2024.131258. URLhttps://www.sciencedirect.com/science/ article/pii/S0360544224010314

    work page doi:10.1016/j.energy.2024.131258 2024

  41. [42]

    Pentsos, S

    V . Pentsos, S. Tragoudas, J. Wibbenmeyer, N. Khdeer, A hybrid lstm-transformer model for power load forecasting, IEEE Transactions on Smart Grid 16 (3) (2025) 2624– 2634.doi:10.1109/TSG.2025.3535407

    work page doi:10.1109/tsg.2025.3535407 2025

  42. [43]

    J. Dong, Y . Jiang, P. Chen, J. Li, Z. Wang, S. Han, Short-term power load forecasting using bidirectional gated recurrent units-based adaptive stacked autoen- coder, International Journal of Electrical Power & Energy Systems 165 (2025) 110459.doi:https: //doi.org/10.1016/j.ijepes.2025.110459. URLhttps://www.sciencedirect.com/science/ article/pii/S0142061...

    work page doi:10.1016/j.ijepes.2025.110459 2025

  43. [44]

    J. Hu, W. Hu, D. Cao, X. Sun, J. Chen, Y . Huang, Z. Chen, F. Blaabjerg, Probabilistic net load fore- casting based on transformer network and gaussian process-enabled residual modeling learning method, Renewable Energy 225 (2024) 120253.doi:https: //doi.org/10.1016/j.renene.2024.120253. URLhttps://www.sciencedirect.com/science/ article/pii/S0960148124003185

    work page doi:10.1016/j.renene.2024.120253 2024

  44. [45]

    Rigotti, C

    M. Rigotti, C. Miksovic, I. Giurgiu, T. Gschwind, P. Scot- ton, Attention-based interpretability with concept trans- formers, in: International conference on learning repre- sentations, 2021

    2021

  45. [46]

    H. Zhao, H. Zhang, C. Tian, Z. Wei, A. Zhou, Srlinear: Lightweight long-term time series forecasting via sym- bolic regression, IEEE Transactions on Artificial Intelli- gence (2025)

    2025

  46. [47]

    X. Han, X. Zhang, Y . Wu, Z. Zhang, Z. Wu, Are kans effective for multivariate time series forecasting? (2025). arXiv:2408.11306. URLhttps://arxiv.org/abs/2408.11306

    arXiv 2025

  47. [48]

    Jiang, Y

    B. Jiang, Y . Wang, Q. Wang, H. Geng, A novel in- terpretable short-term load forecasting method based on kolmogorov-arnold networks, IEEE Transactions on Power Systems 40 (1) (2025) 1180–1183.doi:10.1109/ TPWRS.2024.3498452

    arXiv 2025

  48. [49]

    AIP Advances14(7), 075120 (2024)

    M. Abbas, Y . Che, A. Zafar, Utilizing the new kolmogorov-arnold network for short-term electric load forecasting, Journal of Renewable and Sustainable Energy 17 (2) (2025) 025501.arXiv:https://pubs. aip.org/aip/jrse/article-pdf/doi/10.1063/5. 0253629/20479602/025501\_1\_5.0253629.pdf, doi:10.1063/5.0253629. URLhttps://doi.org/10.1063/5.0253629 22

    work page doi:10.1063/5 2025

  49. [50]

    Jiang, H

    M. Jiang, H. Zhou, D_kan model: Enhancing power load forecasting with kolmogorov-arnold networks, in: 2024 4th International Conference on Energy Engineering and Power Systems (EEPS), 2024, pp. 335–338.doi:10. 1109/EEPS63402.2024.10804420

    arXiv 2024

  50. [51]

    Zhang, L

    Y . Zhang, L. Cui, W. Yan, Integrating kolmogorov–arnold networks with time series prediction framework in electricity demand forecasting, Energies 18 (6) (2025). doi:10.3390/en18061365. URLhttps://www.mdpi.com/1996-1073/18/6/ 1365

    work page doi:10.3390/en18061365 2025

  51. [52]

    M. U. Danish, K. Grolinger, Kolmogorov–arnold recur- rent network for short term load forecasting across diverse consumers, Energy Reports 13 (2025) 713–727.doi: https://doi.org/10.1016/j.egyr.2024.12.038. URLhttps://www.sciencedirect.com/science/ article/pii/S2352484724008539

    work page doi:10.1016/j.egyr.2024.12.038 2025

  52. [53]

    S. Wang, Y . Huang, Q. Wang, Medium-term jointly load forecasting via an enhanced kan-based mtl framework, in: 2024 IEEE PES 16th Asia-Pacific Power and En- ergy Engineering Conference (APPEEC), 2024, pp. 1–5. doi:10.1109/APPEEC61255.2024.10922466

    work page doi:10.1109/appeec61255.2024.10922466 2024

  53. [54]

    Z. Pei, Z. Zhang, J. Chen, W. Liu, B. Chen, Y . Huang, H. Yang, Y . Lu, Kan–cnn: A novel framework for electric vehicle load forecasting with enhanced engineering appli- cability and simplified neural network tuning, Electronics 14 (3) (2025).doi:10.3390/electronics14030414. URLhttps://www.mdpi.com/2079-9292/14/3/414

    work page doi:10.3390/electronics14030414 2025

  54. [55]

    G. Ruan, D. Wu, X. Zheng, H. Zhong, C. Kang, M. Dahleh, S. Sivaranjani, L. Xie, A cross-domain ap- proach to analyzing the short-run impact of covid-19 on the us electricity sector, Joule 4 (11) (2020) 2322–2337. doi:10.1016/j.joule.2020.08.017

    work page doi:10.1016/j.joule.2020.08.017 2020

  55. [56]

    URLhttps://www.google.com/covid19/ mobility/

    Google, Covid-19 community mobility report (2021). URLhttps://www.google.com/covid19/ mobility/

    2021

  56. [57]

    Vaswani, N

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, I. Polosukhin, Attention is all you need, Advances in neural information processing systems 30 (2017)

    2017

  57. [58]

    Noorizadegan, S

    A. Noorizadegan, S. Wang, L. Ling, A practitioner’s guide to kolmogorov-arnold networks (2025).arXiv:2510. 25781. URLhttps://arxiv.org/abs/2510.25781

    Pith/arXiv arXiv 2025

  58. [59]

    C. Zeng, J. Wang, H. Shen, Q. Wang, Kan versus mlp on irregular or noisy functions (2024).arXiv:2408.07906. URLhttps://arxiv.org/abs/2408.07906

    arXiv 2024

  59. [60]

    Y . Cang, Y . hang liu, L. Shi, Can kan work? exploring the potential of kolmogorov-arnold networks in computer vision (2024).arXiv:2411.06727. URLhttps://arxiv.org/abs/2411.06727

    arXiv 2024

  60. [61]

    H. Zhou, S. Zhang, J. Peng, S. Zhang, J. Li, H. Xiong, W. Zhang, Informer: Beyond efficient transformer for long sequence time-series forecasting, in: Proceedings of the AAAI conference on artificial intelligence, V ol. 35, 2021, pp. 11106–11115

    2021

  61. [62]

    A. F. Ansari, L. Stella, C. Turkmen, X. Zhang, P. Mercado, H. Shen, O. Shchur, S. S. Rangapuram, S. P. Arango, S. Kapoor, et al., Chronos: Learning the language of time series, arXiv preprint arXiv:2403.07815 (2024)

    Pith/arXiv arXiv 2024

  62. [63]

    Y . Li, X. Lu, Y . Wang, D. Dou, Generative time series forecasting with diffusion, denoise, and disentanglement, in: S. Koyejo, S. Mohamed, A. Agarwal, D. Belgrave, K. Cho, A. Oh (Eds.), Advances in Neural Information Processing Systems, V ol. 35, Curran Associates, Inc., 2022, pp. 23009–23022. URLhttps://proceedings.neurips. cc/paper_files/paper/2022/fi...

    2022

  63. [64]

    Z. An, J. You, J. Li, Y . Tang, W. Li, H. Du, S. Du, Fredn: Spectral disentanglement for time series forecasting via learnable frequency decomposition (2025).arXiv:2511. 11817. URLhttps://arxiv.org/abs/2511.11817

    arXiv 2025

  64. [65]

    Y . Li, Z. Chen, D. Zha, M. Du, J. Ni, D. Zhang, H. Chen, X. Hu, Towards learning disentangled representations for time series, in: Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD ’22, Association for Computing Machinery, New York, NY , USA, 2022, p. 3270–3278.doi:10.1145/ 3534678.3539140. URLhttps://doi.org/10.11...

    work page doi:10.1145/3534678.3539140 2022

  65. [66]

    Z. Liu, J. Yang, Q. Mao, Y . Zhao, M. Cheng, Z. Li, Q. Liu, E. Chen, Disents: Disentangled channel evolving pattern modeling for multivariate time series forecasting (2024). arXiv:2410.22981. URLhttps://arxiv.org/abs/2410.22981

    arXiv 2024

  66. [67]

    E. N. Cawood, D. Fiorello, P. Christidis, Mobility patterns after the pandemic: a survey in 20 european cities, Transportation Research Procedia 72 (2023) 611–618, tRA Lisbon 2022 Conference Proceedings Transport Research Arena (TRA Lisbon 2022),14th- 17th November 2022, Lisboa, Portugal.doi:https: //doi.org/10.1016/j.trpro.2023.11.446. URLhttps://www.sci...

    work page doi:10.1016/j.trpro.2023.11.446 2023

  67. [68]

    Shokouhyar, S

    S. Shokouhyar, S. Shokoohyar, A. Sobhani, A. J. Gorizi, Shared mobility in post-covid era: New challenges and opportunities, Sustainable Cities and Society 67 (2021) 102714

    2021

  68. [69]

    Zarabi, E

    Z. Zarabi, E. O. D. Waygood, T. Schwanen, Switching back to transit? post-pandemic commute choices and 23 emotional response, Transportation Research Part A: Policy and Practice 203 (2026) 104748.doi:https: //doi.org/10.1016/j.tra.2025.104748. URLhttps://www.sciencedirect.com/science/ article/pii/S0965856425003817 24

    work page doi:10.1016/j.tra.2025.104748 2026