English

POPCORN: Partially Observed Prediction COnstrained ReiNforcement Learning

Machine Learning 2020-04-01 v2 Machine Learning

Abstract

Many medical decision-making tasks can be framed as partially observed Markov decision processes (POMDPs). However, prevailing two-stage approaches that first learn a POMDP and then solve it often fail because the model that best fits the data may not be well suited for planning. We introduce a new optimization objective that (a) produces both high-performing policies and high-quality generative models, even when some observations are irrelevant for planning, and (b) does so in batch off-policy settings that are typical in healthcare, when only retrospective data is available. We demonstrate our approach on synthetic examples and a challenging medical decision-making problem.

Keywords

Cite

@article{arxiv.2001.04032,
  title  = {POPCORN: Partially Observed Prediction COnstrained ReiNforcement Learning},
  author = {Joseph Futoma and Michael C. Hughes and Finale Doshi-Velez},
  journal= {arXiv preprint arXiv:2001.04032},
  year   = {2020}
}

Comments

Accepted to AISTATS 2020, Palermo, Italy

R2 v1 2026-06-23T13:09:12.271Z