mlsanityData to 14 Oct 2026

Learning-to-Defer in Non-Stationary Time Series via Switching State-Space Models

arXiv 2601.22538 ↗First posted 30 Jan 2026cs.LG stat.APcited by 6

Not on the map. Why?

Abstract (arXiv)

Learning-to-defer (L2D) lets a predictor decide, at each round, whether to issue its own forecast or pay for an expert's. In non-stationary time series this decision must keep adapting, although deployment reveals only the consulted expert's forecast while a historical archive records every expert with the target. L2D-SLDS learns from this archive a switching state-space model of the target and all expert forecasts, whose shared and expert-specific states describe how experts move together and apart. Its predictive law supplies the internal forecast and the expected cost of every consultation, which a greedy router minimizes, and one consultation also updates the beliefs about unconsulted and unavailable experts. We prove sublinear regret against a changing conditional-risk oracle without exploration, when the candidate models are accurate and either the archive separates them or live feedback reveals cost differences. On three real datasets, L2D-SLDS has the lowest cost among eight bandit routers and adapts its consultation rate to the fee.

Key facts. Cited by 6 papers in the map (data through 14 Oct 2026). Builds on 36 of its 112 references that are in the map. First posted on arXiv 30 Jan 2026.

Builds on

36 of 112 references.

References · 36

  1. 2017Attention Is All You Needcited by 60,361
  2. 2017Selective Classification for Deep Neural Networkscited by 852
  3. 2019Financial Time Series Forecasting with Deep Learning : A Systematic Literature Review: 2005-2019cited by 273
  4. 2020Consistent Estimators for Learning to Defer to an Expertcited by 271
  5. 2017Predict Responsibly: Improving Fairness and Accuracy by Learning to Defercited by 233
  6. 2020Beyond UCB: Optimal and Efficient Contextual Bandits with Regression Oraclescited by 197
  7. 2014Learning to Optimize via Information-Directed Samplingcited by 168
  8. 2019Weighted Linear Bandits for Non-Stationary Environmentscited by 86
  9. 2017Mostly Exploration-Free Algorithms for Contextual Banditscited by 85
  10. 2015Online Learning with Feedback Graphs: Beyond Banditscited by 80
  11. 2023Who Should Predict? Exact Algorithms For Learning to Defer to Humanscited by 70
  12. 2017Ensemble Samplingcited by 62
24 more
  1. 2022Learning to Defer to Multiple Experts: Consistent Surrogate Losses, Confidence Calibration, and Conformal Ensemblescited by 60
  2. 2017A Change-Detection based Framework for Piecewise-stationary Multi-Armed Bandit Problemcited by 49
  3. 2022Sample Efficient Learning of Predictors that Complement Humanscited by 46
  4. 2023Principled Approaches for Learning to Defer with Multiple Expertscited by 38
  5. 2016Recurrent switching linear dynamical systemscited by 35
  6. 2024Regression with Multi-Expert Deferralcited by 29
  7. 2023In Defense of Softmax Parametrization for Calibrated and Consistent Learning to Defercited by 26
  8. 2024Realizable H-Consistent and Bayes-Consistent Loss Functions for Learning to Defercited by 26
  9. 2024Cardinality-Aware Set Prediction and Top-k Classificationcited by 23
  10. 2014Particle Gibbs with Ancestor Samplingcited by 21
  11. 2024A Two-Stage Learning-to-Defer Approach for Multi-Task Learningcited by 17
  12. 2025Adversarial Robustness in Two-Stage Learning-to-Defer: Algorithms and Guaranteescited by 16
  13. 2025Mastering Multiple-Expert Routing: Realizable H-Consistency and Strong Guarantees for Learning to Defercited by 16
  14. 2024Optimal Query Allocation in Extractive QA with LLMs: A Learning-to-Defer Framework with Theoretical Guaranteescited by 15
  15. 2024A Causal Framework for Evaluating Deferring Systemscited by 13
  16. 2025Budgeted Multiple-Expert Deferralcited by 13
  17. 2021Learning-to-defer for sequential medical decision-making under uncertaintycited by 12
  18. 2023An Information-Theoretic Analysis of Nonstationary Bandit Learningcited by 11
  19. 2024Modeling Latent Neural Dynamics with Gaussian Process Switching Linear Dynamical Systemscited by 11
  20. 2021Neural HMMs are all you need (for high-quality attention-free TTS)cited by 8
  21. 2025Theory and Algorithms for Learning with Multi-Class Abstention and Multi-Expert Deferralcited by 8
  22. 2016Dynamic Bayesian Predictive Synthesis in Time Series Forecastingcited by 7
  23. 2025Why Ask One When You Can Ask k? Learning-to-Defer to the Top-k Expertscited by 7
  24. 2025Adversarial Robustness in One-Stage Learning-to-Defercited by 5

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 6.

6 papers.

  1. 2026Beyond Augmented-Action Surrogates for Multi-Expert Learning-to-Defercited by 5
  2. 2026Consistent Learning-to-Defer with Expert-Conditional Advicecited by 3
  3. 2026Optimized Deferral for Imbalanced Settingscited by 2
  4. 2026Online Learning-to-Defer with Varying Expertscited by 1
  5. 2026Mind the Gap: Structure-Aware Consistency in Preference Learningcited by 0
  6. 2026A Query Is Not a Commitment: Learning to Correct Expert Answers in Online Deferralcited by 0