On Overfitting and Asymptotic Bias in Batch Reinforcement Learning with Partial Observability (Extended Abstract)

Vincent Francois-Lavet; Guillaume Rabusseau; Joelle Pineau; Damien Ernst; Raphael Fonteneau

IJCAI 2020poster0 citations

On Overfitting and Asymptotic Bias in Batch Reinforcement Learning with Partial Observability (Extended Abstract)

Vincent Francois-Lavet, Guillaume Rabusseau, Joelle Pineau, Damien Ernst, Raphael Fonteneau

Abstract

When an agent has limited information on its environment, the suboptimality of an RL algorithm can be decomposed into the sum of two terms: a term related to an asymptotic bias (suboptimality with unlimited data) and a term due to overfitting (additional suboptimality due to limited data). In the context of reinforcement learning with partial observability, this paper provides an analysis of the tradeoff between these two error sources. In particular, our theoretical analysis formally characterizes how a smaller state representation increases the asymptotic bias while decreasing the risk of overfitting.

Machine Learning: Reinforcement LearningPlanning and Scheduling: POMDPsKnowledge Representation and Reasoning: Reasoning about Knowledge and BeliefMachine Learning: Learning Theory

BibTeX

@inproceedings{ijcai2020p706,
  title     = {On Overfitting and Asymptotic Bias in Batch Reinforcement Learning with Partial Observability (Extended Abstract)},
  author    = {Francois-Lavet, Vincent and Rabusseau, Guillaume and Pineau, Joelle and Ernst, Damien and Fonteneau, Raphael},
  booktitle = {Proceedings of the Twenty-Ninth International Joint Conference on
               Artificial Intelligence, {IJCAI-20}},
  publisher = {International Joint Conferences on Artificial Intelligence Organization},
  editor    = {Christian Bessiere},
  pages     = {5055--5059},
  year      = {2020},
  month     = {7},
  note      = {Journal track},
  doi       = {10.24963/ijcai.2020/706},
  url       = {https://doi.org/10.24963/ijcai.2020/706},
}