LGMLJun 2, 2019

On the Correctness and Sample Complexity of Inverse Reinforcement Learning

arXiv:1906.00422v115 citations
Originality Incremental advance
AI Analysis

This work addresses the theoretical correctness and efficiency of IRL for finite-state Markov Decision Processes, providing a rigorous sample complexity bound, though it is incremental as it builds on existing IRL frameworks.

The paper tackles the problem of inverse reinforcement learning (IRL) by proposing an L1-regularized SVM formulation based on a geometric analysis, and shows a sample complexity of O(n^2 log(nk)) for recovering a reward function that generates a policy satisfying Bellman's optimality condition.

Inverse reinforcement learning (IRL) is the problem of finding a reward function that generates a given optimal policy for a given Markov Decision Process. This paper looks at an algorithmic-independent geometric analysis of the IRL problem with finite states and actions. A L1-regularized Support Vector Machine formulation of the IRL problem motivated by the geometric analysis is then proposed with the basic objective of the inverse reinforcement problem in mind: to find a reward function that generates a specified optimal policy. The paper further analyzes the proposed formulation of inverse reinforcement learning with $n$ states and $k$ actions, and shows a sample complexity of $O(n^2 \log (nk))$ for recovering a reward function that generates a policy that satisfies Bellman's optimality condition with respect to the true transition probabilities.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes