Learning Control by Iterative Inversion

Gal Leibovich, Guy Jacob, Or Avner, Gal Novik, Aviv Tamar

Research output: Contribution to journalConference articlepeer-review

Abstract

We propose iterative inversion - an algorithm for learning an inverse function without input-output pairs, but only with samples from the desired output distribution and access to the forward function. The key challenge is a distribution shift between the desired outputs and the outputs of an initial random guess, and we prove that iterative inversion can steer the learning correctly, under rather strict conditions on the function. We apply iterative inversion to learn control. Our input is a set of demonstrations of desired behavior, given as video embeddings of trajectories (without actions), and our method iteratively learns to imitate trajectories generated by the current policy, perturbed by random exploration noise. Our approach does not require rewards, and only employs supervised learning, which can be easily scaled to use state-of-the-art trajectory embedding techniques and policy representations. Indeed, with a VQ-VAE embedding, and a transformer-based policy, we demonstrate nontrivial continuous control on several tasks (videos available at https://sites.google.com/view/iter-inver). Further, we report an improved performance on imitating diverse behaviors compared to reward based methods.

Original languageEnglish
Pages (from-to)19200-19227
Number of pages28
JournalProceedings of Machine Learning Research
Volume202
StatePublished - 2023
Event40th International Conference on Machine Learning, ICML 2023 - Honolulu, United States
Duration: 23 Jul 202329 Jul 2023

ASJC Scopus subject areas

  • Artificial Intelligence
  • Software
  • Control and Systems Engineering
  • Statistics and Probability

Fingerprint

Dive into the research topics of 'Learning Control by Iterative Inversion'. Together they form a unique fingerprint.

Cite this