mlsanityData to 14 Oct 2026

Optimal Empirical Risk Minimization under Temporal Distribution Shifts

arXiv 2507.13287 ↗First posted 17 Jul 2025stat.ME cs.LGcited by 2

Not on the map. Why?

Abstract (arXiv)

Temporal distribution shifts pose a key challenge for machine learning models trained and deployed in dynamically evolving environments. This paper introduces RIDER (RIsk minimization under Dynamically Evolving Regimes) which derives optimally-weighted empirical risk minimization procedures under temporal distribution shifts. Our approach is theoretically grounded in the random distribution shift model, where random shifts arise as a superposition of numerous unpredictable changes in the data-generating process. We show that common weighting schemes, such as pooling all data, exponentially weighting data, and using only the most recent data, emerge naturally as special cases in our framework. We demonstrate that RIDER consistently improves out-of-sample predictive performance when applied as a fine-tuning step on the Yearbook dataset, across a range of benchmark methods in Wild-Time. Moreover, we show that RIDER outperforms standard weighting strategies in two other real-world tasks: predicting stock market volatility and forecasting ride durations in NYC taxi data.

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 36 of its 62 references that are in the map. First posted on arXiv 17 Jul 2025.

Builds on

36 of 62 references.

References · 36

  1. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 12,049
  2. 2016Overcoming catastrophic forgetting in neural networkscited by 4,787
  3. 2020Unsupervised Learning of Visual Features by Contrasting Cluster Assignmentscited by 2,857
  4. 2016iCaRL: Incremental Classifier and Representation Learningcited by 2,333
  5. 2014Unsupervised Domain Adaptation by Backpropagationcited by 1,917
  6. 2017Gradient Episodic Memory for Continual Learningcited by 1,898
  7. 2019Invariant Risk Minimizationcited by 1,854
  8. 2020WILDS: A Benchmark of in-the-Wild Distribution Shiftscited by 1,375
  9. 2019Distributionally Robust Neural Networks for Group Shifts: On the Importance of Regularization for Worst-Case Generalizationcited by 1,247
  10. 2018Averaging Weights Leads to Wider Optima and Better Generalizationcited by 1,209
  11. 2016Deep CORAL: Correlation Alignment for Deep Domain Adaptationcited by 1,200
  12. 2018Efficient Lifelong Learning with A-GEMcited by 1,058
24 more
  1. 2020In Search of Lost Domain Generalizationcited by 1,054
  2. 2018Progress & Compress: A scalable framework for continual learningcited by 520
  3. 2017Functional Map of the Worldcited by 338
  4. 2018Learning Models with Uniform Performance via Distributionally Robust Optimizationcited by 335
  5. 2021Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Developmentcited by 264
  6. 2021Mind the Gap: Assessing Temporal Generalization in Neural Language Modelscited by 230
  7. 2019On the Use of ArXiv as a Datasetcited by 103
  8. 2022Wild-Time: A Benchmark of in-the-Wild Distribution Shift over Timecited by 90
  9. 2015A Century of Portraits: A Visual Historical Record of American High School Yearbookscited by 54
  10. 2019Continual Learning with Adaptive Weights (CLAW)cited by 38
  11. 2022Temporal Domain Generalization with Drift-Aware Dynamic Neural Networkscited by 33
  12. 2021Training for the Future: A Simple Gradient Interpolation Loss to Generalize Along Timecited by 28
  13. 2022Generalizing to Evolving Domains with Latent Structure-Aware Sequential Autoencodercited by 19
  14. 2023An Adaptive Algorithm for Learning with Unknown Distribution Driftcited by 16
  15. 2024Generalizing across Temporal Domains with Koopman Operatorscited by 13
  16. 2015Statistical Learning under Nonstationary Mixing Processescited by 12
  17. 2023Learning under random distributional shiftscited by 6
  18. 2024Enhancing Evolving Domain Generalization through Dynamic Latent Representationscited by 6
  19. 2022Time Series Prediction under Distribution Shift using Differentiable Forgettingcited by 5
  20. 2024Out-of-distribution generalization under random, dense distributional shiftscited by 5
  21. 2024Exponentially Weighted Moving Modelscited by 4
  22. 2024Non-stationary Domain Generalization: Theory and Algorithmcited by 4
  23. 2024Beyond Reweighting: On the Predictive Role of Covariate Shift in Effect Generalizationcited by 4
  24. 2022BAM: Bayes with Adaptive Memorycited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026Fast Rates for Nonstationary Weighted Risk Minimizationcited by 1
  2. 2026Beyond Discrepancy: Oracle Drift in Nonstationary Weighted Risk Minimizationcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Optimizing Return Distributions with Distributional Dynamic Programming
  2. 2025Dynamic data generation and dynamic portfolio selection: an application of a score-based diffusion model
  3. 2026EvoTS-Agent: A Self-Evolving LLM Agent for Financial Time Series Change Point Detection
  4. 2025Dynamic Pricing in the Linear Valuation Model using Shape Constraints
  5. 2025Multi-period Asset-liability Management with Reinforcement Learning in a Regime-Switching Market
  6. 2026Addressing Market Regime Changes and Heavy-Tailed Returns in Portfolio Optimization via Bayesian VAR and Elliptical Black-Litterman
  7. 2023Risk Aware Benchmarking of Large Language Models
  8. 2026Portfolio Reinforcement Learning with Scenario-Context Rollout