Chapter 16: Practical reinforcement learning in dynamic treatment regimes

Robert D. Vincent, Joëlle Pineau, Norma Ybarra, Issam El NaqaSociety 2015 · DOI · OpenAlex

Abstract

16.1 ▪ IntroductionReinforcement learning (RL) is a general computational approach for optimizing a sequence of actions for a (possibly unknown) dynamic system. In this chapter we propose a framework for applying RL algorithms to derive optimal dynamic treatment regimes. The goal of the RL approach is to discover dynamic treatment regimes directly from data, perhaps taken from clinical trials, laboratory experiments, or computational models. The power of the method lies in its ability to do this optimization without requiring models of the underlying disease processes, while still permitting the incorporation of insights learned from such a model, if available.
BibTeX
@inproceedings{vincent2015chapter,
  author = {Robert D. Vincent and Joëlle Pineau and Norma Ybarra and Issam El Naqa},
  title = {{Chapter 16: Practical reinforcement learning in dynamic treatment regimes}},
  booktitle = {Society},
  year = {2015},
  doi = {10.1137/1.9781611974188.ch16},
}

Loading related content...

Figure