Compatible Reward Inverse Reinforcement Learning - Archive ouverte HAL Access content directly
Conference Papers Year :

Compatible Reward Inverse Reinforcement Learning

(1) , (2) , (1)
1
2

Abstract

Inverse Reinforcement Learning (IRL) is an effective approach to recover a reward function that explains the behavior of an expert by observing a set of demonstrations. This paper is about a novel model-free IRL approach that, differently from most of the existing IRL algorithms, does not require to specify a function space where to search for the expert's reward function. Leveraging on the fact that the policy gradient needs to be zero for any optimal policy, the algorithm generates a set of basis functions that span the subspace of reward functions that make the policy gradient vanish. Within this subspace, using a second-order criterion, we search for the reward function that penalizes the most a deviation from the expert's policy. After introducing our approach for finite domains, we extend it to continuous ones. The proposed approach is empirically compared to other IRL methods both in the (finite) Taxi domain and in the (continuous) Linear Quadratic Gaussian (LQG) and Car on the Hill environments.
Fichier principal
Vignette du fichier
6800-compatible-reward-inverse-reinforcement-learning.pdf (452.54 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-01653328 , version 1 (01-12-2017)

Identifiers

  • HAL Id : hal-01653328 , version 1

Cite

Alberto Maria Metelli, Matteo Pirotta, Marcello Restelli. Compatible Reward Inverse Reinforcement Learning. The Thirty-first Annual Conference on Neural Information Processing Systems - NIPS 2017, Dec 2017, Long Beach, United States. ⟨hal-01653328⟩
203 View
240 Download

Share

Gmail Facebook Twitter LinkedIn More