mlsanityData to 7 Oct 2026

Explaining Time Series Forecasting with Horizon-Resolved Attribution

arXiv 2609.12639 ↗First posted 11 Sep 2026cs.LG cs.AIcited by 0

Not on the map. Why?

Abstract (arXiv)

Recent advances in explaining time series (TS) models have produced methods that identify which past values a prediction depends on. However, most existing methods return a single importance vector, assuming that every predicted step depends on the same past values. In this paper, we show that this assumption does not hold, as different forecast steps depend on different past values. Motivated by this observation, we propose Horizon-Resolved eXplanation (HRX), which adds a horizon axis to the explanation, so that every forecast step receives its own importance map. HRX is a simple yet effective plug-in framework with three components: 1) an estimator that reads these maps out of any differentiable forecaster without modifying the TS backbone, 2) an evaluation protocol that validates the horizon axis by measuring how much a single forecast step changes when the inputs an importance map ranks highest are removed, and 3) a rank criterion that predicts in advance whether the axis is worth resolving on a given TS. We further show that this step-wise dependence is low-dimensional, as the explanations of all steps are built from a few shared maps whose number does not grow with the forecast length. Extensive experiments across various backbones and datasets show that the improvement comes from the horizon axis and holds for estimators of previous explanation methods. Code is available at https://github.com/seunghan96/HRX.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 58 of its 65 references that are in the map. First posted on arXiv 11 Sep 2026.

Builds on

58 of 65 references.

References · 58

  1. 2017Attention Is All You Needcited by 59,932
  2. 2017A Unified Approach to Interpreting Model Predictionscited by 6,488
  3. 2017Axiomatic Attribution for Deep Networkscited by 3,350
  4. 2013Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Mapscited by 2,348
  5. 2020Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecastingcited by 2,141
  6. 2022A Time Series is Worth 64 Words: Long-term Forecasting with Transformerscited by 1,981
  7. 2020Reformer: The Efficient Transformercited by 1,658
  8. 2021Autoformer: Decomposition Transformers with Auto-Correlation for Long-Term Series Forecastingcited by 1,654
  9. 2022Are Transformers Effective for Time Series Forecasting?cited by 1,400
  10. 2022FEDformer: Frequency Enhanced Decomposed Transformer for Long-term Series Forecastingcited by 1,200
  11. 2018Sanity Checks for Saliency Mapscited by 1,149
  12. 2017SmoothGrad: removing noise by adding noisecited by 1,097
38 more
  1. 2022TimesNet: Temporal 2D-Variation Modeling for General Time Series Analysiscited by 1,030
  2. 2023iTransformer: Inverted Transformers Are Effective for Time Series Forecastingcited by 906
  3. 2018RISE: Randomized Input Sampling for Explanation of Black-box Modelscited by 825
  4. 2019Temporal Fusion Transformers for Interpretable Multi-horizon Time Series Forecastingcited by 748
  5. 2019N-BEATS: Neural basis expansion analysis for interpretable time series forecastingcited by 744
  6. 2024Chronos: Learning the Language of Time Seriescited by 712
  7. 2023A decoder-only foundation model for time-series forecastingcited by 650
  8. 2024Unified Training of Universal Time Series Forecasting Transformerscited by 588
  9. 2019ERASER: A Benchmark to Evaluate Rationalized NLP Modelscited by 523
  10. 2018A Benchmark for Interpretability Methods in Deep Neural Networkscited by 467
  11. 2024MOMENT: A Family of Open Time-series Foundation Modelscited by 453
  12. 2015Evaluating the visualization of what a Deep Neural Network has learnedcited by 416
  13. 2022Non-stationary Transformers: Exploring the Stationarity in Time Series Forecastingcited by 385
  14. 2021SCINet: Time Series Modeling and Forecasting with Sample Convolution and Interactioncited by 335
  15. 2024TimeMixer: Decomposable Multiscale Mixing for Time Series Forecastingcited by 333
  16. 2023Long-term Forecasting with TiDE: Time-series Dense Encodercited by 327
  17. 2024Time-MoE: Billion-Scale Time Series Foundation Models with Mixture of Expertscited by 280
  18. 2024TimeXer: Empowering Transformers for Time Series Forecasting with Exogenous Variablescited by 241
  19. 2023TSMixer: An All-MLP Architecture for Time Series Forecastingcited by 232
  20. 2024Timer: Generative Pre-trained Transformers Are Large Time Series Modelscited by 198
  21. 2024Deep Time Series Models: A Comprehensive Survey and Benchmarkcited by 191
  22. 2024GIFT-Eval: A Benchmark For General Time Series Forecasting Model Evaluationcited by 188
  23. 2023Frequency-domain MLPs are More Effective Learners in Time Series Forecastingcited by 170
  24. 2025Sundial: A Family of Highly Capable Time Series Foundation Modelscited by 168
  25. 2022Less Is More: Fast Multivariate Time Series Forecasting with Light Sampling-oriented MLP Structurescited by 158
  26. 2022FiLM: Frequency improved Legendre Memory Model for Long-term Time Series Forecastingcited by 147
  27. 2024Tiny Time Mixers (TTMs): Fast Pre-trained Models for Enhanced Zero/Few-Shot Forecasting of Multivariate Time Seriescited by 120
  28. 2020Benchmarking Deep Learning Interpretability in Time Series Predictionscited by 102
  29. 2024VisionTS: Visual Masked Autoencoders Are Free-Lunch Zero-Shot Time Series Forecasterscited by 87
  30. 2020TimeSHAP: Explaining Recurrent Models through Sequence Perturbationscited by 73
  31. 2021Explaining Time Series Predictions with Dynamic Maskscited by 72
  32. 2023MSGNet: Learning Multi-Scale Inter-Series Correlations for Multivariate Time Series Forecastingcited by 50
  33. 2023Encoding Time-Series Explanations through Self-Supervised Model Behavior Consistencycited by 42
  34. 2023Learning to Embed Time Series Patches Independentlycited by 34
  35. 2024TimeX++: Learning Time-Series Explanations with Information Bottleneckcited by 32
  36. 2021Temporal Dependencies in Feature Importance for Time Series Predictionscited by 27
  37. 2023Learning Perturbations to Explain Time Series Predictionscited by 23
  38. 2024Explaining Time Series via Contrastive and Locally Sparse Perturbationscited by 22

50 most cited of 58 All 58 (JSON)

Built on by

Not cited yet.