mlsanityData to 14 Oct 2026

Bayesian Safe Policy Learning with Chance Constrained Optimization: Application to Military Security Assessment during the Vietnam War

arXiv 2307.08840 ↗First posted 17 Jul 2023cs.LG stat.APcited by 4

Not on the map. Why?

Abstract (arXiv)

Algorithmic decisions and recommendations are used in many high-stakes decision-making settings such as criminal justice, medicine, and public policy. We investigate whether it would have been possible to improve a security assessment algorithm employed during the Vietnam War, using outcomes measured immediately after its introduction in late 1969. This empirical application raises several methodological challenges that frequently arise in high-stakes algorithmic decision-making. First, before implementing a new algorithm, it is essential to characterize and control the risk of yielding worse outcomes than the existing algorithm. Second, the existing algorithm is deterministic, and learning a new algorithm requires transparent extrapolation. Third, the existing algorithm involves discrete decision tables that are difficult to optimize over. To address these challenges, we introduce the Average Conditional Risk (ACRisk), which first quantifies the risk that a new algorithmic policy leads to worse outcomes for subgroups of individual units and then averages this over the distribution of subgroups. We also propose a Bayesian policy learning framework that maximizes the posterior expected value while controlling the posterior expected ACRisk. This framework separates the estimation of heterogeneous treatment effects from policy optimization, enabling flexible estimation of effects and optimization over complex policy classes. We characterize the resulting chance-constrained optimization problem as a constrained linear programming problem. Our analysis shows that compared to the actual algorithm used during the Vietnam War, the learned algorithm assesses most regions as more secure and emphasizes economic and political factors over military factors.

Key facts. Cited by 4 papers in the map (data through 14 Oct 2026). Builds on 31 of its 87 references that are in the map. First posted on arXiv 17 Jul 2023.

Builds on

31 of 87 references.

References · 31

  1. 2020Is Pessimism Provably Efficient for Offline RL?cited by 364
  2. 2021Bridging Offline Reinforcement Learning and Imitation Learning: A Tale of Pessimismcited by 279
  3. 2021Bellman-consistent Pessimism for Offline Reinforcement Learningcited by 269
  4. 2021Pessimistic Model-based Offline Reinforcement Learning under Partial Coveragecited by 164
  5. 2022Pessimistic Bootstrapping for Uncertainty-Driven Offline Reinforcement Learningcited by 125
  6. 2021Provable Benefits of Actor-Critic Methods for Offline Reinforcement Learningcited by 110
  7. 2020The Importance of Pessimism in Fixed-Dataset Policy Optimizationcited by 99
  8. 2022Pessimistic Q-Learning for Offline Reinforcement Learning: Towards Optimal Sample Complexitycited by 85
  9. 2021Towards Instance-Optimal Offline Reinforcement Learning with Pessimismcited by 78
  10. 2021Discussion on "Bayesian Regression Tree Models for Causal Inference: Regularization, Confounding, and Heterogeneous Effects" by Hahn, Murray and Carvalhocited by 77
  11. 2018Offline Multi-Action Policy Learning: Generalization and Optimizationcited by 65
  12. 2017Balanced Policy Evaluation and Learningcited by 52
19 more
  1. 2018A Simple and Effective Model-Based Variable Importance Measurecited by 51
  2. 2020Algorithmic Fairness in Educationcited by 44
  3. 2022The Efficacy of Pessimism in Asynchronous Q-Learningcited by 42
  4. 2015Residual Weighted Learning for Estimating Individualized Treatment Rulescited by 38
  5. 2022Offline Reinforcement Learning Under Value and Density-Ratio Realizability: The Power of Gapscited by 34
  6. 2022Policy learning "without" overlap: Pessimism and generalized empirical Bernstein's inequalitycited by 32
  7. 2022Bayesian Causal Inference: A Critical Reviewcited by 21
  8. 2020Experimental Evaluation of Algorithm-Assisted Human Decision-Making: Application to Pretrial Public Safety Assessmentcited by 18
  9. 2020Estimating Optimal Treatment Rules with an Instrumental Variable: A Partial Identification Learning Approachcited by 17
  10. 2017Causal Inference: A Missing Data Perspectivecited by 15
  11. 2022What's the Harm? Sharp Bounds on the Fraction Negatively Affected by Treatmentcited by 15
  12. 2021Safe Chance Constrained Reinforcement Learning for Batch Process Controlcited by 11
  13. 2021Safe Policy Learning through Extrapolation: Application to Pre-trial Risk Assessmentcited by 10
  14. 2022Policy Learning with Asymmetric Counterfactual Utilitiescited by 9
  15. 2022On the Safety of Interpretable Machine Learning: A Maximum Deviation Approachcited by 7
  16. 2022Unit Selection: Learning Benefit Function from Finite Population Datacited by 6
  17. 2017A Nonparametric Bayesian Methodology for Regression Discontinuity Designscited by 5
  18. 2021Optimal Decision Rules Under Partial Identificationcited by 3
  19. 2022Safe Policy Learning under Regression Discontinuity Designs with Multiple Cutoffscited by 3

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2023: 2; 2024: 1; 2025: 0; 2026: 1.

4 papers.

  1. 2023Positivity-free Policy Learning with Observational Datacited by 6
  2. 2023Individualized Policy Evaluation and Learning under Clustered Network Interferencecited by 6
  3. 2024What Sketch Explainability Really Means for Downstream Taskscited by 4
  4. 2026Conformal Risk-Averse Decision Making with Action Conditional Guaranteecited by 2