mlsanityData to 14 Oct 2026

Reinforcement Learning for Durable Algorithmic Recourse

arXiv 2509.22102 ↗First posted 26 Sep 2025cs.LG cs.AIcited by 3

On the map: Deep reinforcement learning methods → RL for games and combinatorial optimization

Abstract (arXiv)

Algorithmic recourse seeks to provide individuals with actionable recommendations that increase their chances of receiving favorable outcomes from automated decision systems (e.g., loan approvals). While prior research has emphasized robustness to model updates, considerably less attention has been given to the temporal dynamics of recourse--particularly in competitive, resource-constrained settings where recommendations shape future applicant pools. In this work, we present a novel time-aware framework for algorithmic recourse, explicitly modeling how candidate populations adapt in response to recommendations. Additionally, we introduce a novel reinforcement learning (RL)-based recourse algorithm that captures the evolving dynamics of the environment to generate recommendations that are both feasible and valid. We design our recommendations to be durable, supporting validity over a predefined time horizon T. This durability allows individuals to confidently reapply after taking time to implement the suggested changes. Through extensive experiments in complex simulation environments, we show that our approach substantially outperforms existing baselines, offering a superior balance between feasibility and long-term validity. Together, these results underscore the importance of incorporating temporal and behavioral dynamics into the design of practical recourse systems.

Key facts. Cited by 3 papers in the map (data through 14 Oct 2026). Placed in RL for games and combinatorial optimization, one of 15 directions in Deep reinforcement learning methods. Builds on 26 of its 39 references that are in the map. First posted on arXiv 26 Sep 2025.

Builds on

26 of 39 references.

References · 26

  1. 2018Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actorcited by 4,885
  2. 2017Counterfactual Explanations without Opening the Black Box: Automated Decisions and the GDPRcited by 1,089
  3. 2019Explaining Machine Learning Classifiers through Diverse Counterfactual Explanationscited by 614
  4. 2018Actionable Recourse in Linear Classificationcited by 328
  5. 2020Algorithmic Recourse: from Counterfactual Explanations to Interventionscited by 250
  6. 2020Multi-Objective Counterfactual Explanationscited by 157
  7. 2020Algorithmic recourse under imperfect causal knowledge: a probabilistic approachcited by 140
  8. 2019The Hidden Assumptions Behind Counterfactual Explanations and Principal Reasonscited by 115
  9. 2021Towards Robust and Reliable Algorithmic Recoursecited by 102
  10. 2021On the Adversarial Robustness of Causal Algorithmic Recoursecited by 56
  11. 2022Probabilistically Robust Recourse: Navigating the Trade-offs between Costs and Robustness in Algorithmic Recoursecited by 40
  12. 2020Algorithmic Recourse in the Wild: Understanding the Impact of Data and Model Shiftscited by 36
14 more
  1. 2022On the Trade-Off between Actionable Explanations and the Right to be Forgottencited by 20
  2. 2022Improvement-Focused Causal Recourse (ICR)cited by 20
  3. 2021CARE: Coherent Actionable Recourse based on Sound Counterfactual Explanationscited by 17
  4. 2023Generating robust counterfactual explanationscited by 13
  5. 2023Setting the Right Expectations: Algorithmic Recourse Over Timecited by 13
  6. 2023Distributionally Robust Recourse Actioncited by 12
  7. 2024Counterfactual Explanations with Probabilistic Guarantees on their Robustness to Model Changecited by 10
  8. 2022Personalized Algorithmic Recourse with Preference Elicitationcited by 9
  9. 2023Endogenous Macrodynamics in Algorithmic Recoursecited by 9
  10. 2024Time Can Invalidate Algorithmic Recoursecited by 8
  11. 2024Feature Responsiveness Scores: Model-Agnostic Explanations for Recoursecited by 8
  12. 2023The Importance of Time in Causal Algorithmic Recoursecited by 2
  13. 2024Perfect Counterfactuals in Imperfect Worlds: Modelling Noisy Implementation of Actions in Sequential Algorithmic Recoursecited by 2
  14. 2023SafeAR: Safe Algorithmic Recourse by Risk-Aware Policiescited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 3.

3 papers.

  1. 2026ElliCE: Efficient and Provably Robust Algorithmic Recourse via the Rashomon Setscited by 2
  2. 2026From Universal to Individualized Actionability: Revisiting Personalization in Algorithmic Recoursecited by 1
  3. 2026Algorithmic Recourse Under Competitioncited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Time-Constrained Recommendations: Reinforcement Learning Strategies for E-Commerce
  2. 2024Deep Reinforcement Learning for Dynamic Algorithm Selection: A Proof-of-Principle Study on Differential Evolution
  3. 2021Accelerating Offline Reinforcement Learning Application in Real-Time Bidding and Recommendation: Potential Use of Simulation
  4. 2025Closing the Loop: Coordinating Inventory and Recommendation via Deep Reinforcement Learning on Multiple Timescales
  5. 2025Do You Trust the Process?: Modeling Institutional Trust for Community Adoption of Reinforcement Learning Policies
  6. 2026Reinforcement Learning for Micro-Level Claims Reserving
  7. 2026Scalable In-Context Reinforcement Learning with Recurrent Algorithm Distillation
  8. 2024Using deep reinforcement learning to promote sustainable human behaviour on a common pool resource problem