Of Moments and Matching: A Game-Theoretic Framework for Closing the Imitation Gap
By Gokul Swamy, Sanjiban Choudhury, J. Bagnell, Steven Wu. 2021
95 citations recorded by the source corpus.
We provide a unifying view of a large family of previous imitation learning algorithms through the lens of moment matching. At its core, our classification scheme is based on whether the learner attempts to match (1) reward or (2) action-value moments of the expert's behavior, with each option leading to differing algorithmic approaches. By considering adversarially chosen divergences between learner and expert behavior, we are able to derive bounds on policy performance that apply for all algorithms in each of these classes, the first to our knowledge. We also introduce the notion of moment recoverability, implicit in many previous analyses of imitation learning, which allows us to cleanly delineate how well each algorithmic family is able to mitigate compounding errors. We derive three novel algorithm templates (AdVIL, AdRIL, and DAeQuIL) with strong guarantees, simple implementation, and competitive empirical performance.
Subfields: Reinforcement Learning
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Reinforcement Learning
- Hannaneh Hajishirzi
- Yejin Choi
- Kianté Brantley
- Christian Bauckhage
- Prithviraj Ammanabrolu
- Rafet Sifa
- Rajkumar Ramamurthy
- Jack Hessel
Citations in the corpus
- Is Reinforcement Learning (Not) for Natural Language Processing?: Benchmarks, Baselines, and Building Blocks for Natural Language Policy Optimization (2022, background): “, 2022) and the cliff MDP problem (Huszár, 2015; Agarwal et al., 2019; Swamy et al., 2021).”