Stronger Generalization Guarantees for Robot Learning by Combining Generative Models and Real-World Data

Abhinav Agarwal, Sushant Veer, Allen Z. Ren, Anirudha Majumdar

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

We are motivated by the problem of learning policies for robotic systems with rich sensory inputs (e.g., vision) in a manner that allows us to guarantee generalization to environments unseen during training. We provide a framework for providing such generalization guarantees by leveraging a finite dataset of real-world environments in combination with a (potentially inaccurate) generative model of environments. The key idea behind our approach is to utilize the generative model in order to implicitly specify a prior over policies. This prior is updated using the real-world dataset of environments by minimizing an upper bound on the expected cost across novel environments derived via Probably Approximately Correct (PAC)-Bayes generalization theory. We demonstrate our approach on two simulated systems with nonlinear/hybrid dynamics and rich sensing modalities: (i) quadrotor navigation with an onboard vision sensor, and (ii) grasping objects using a depth sensor. Comparisons with prior work demonstrate the ability of our approach to obtain stronger generalization guarantees by utilizing generative models. We also present hardware experiments for validating our bounds for the grasping task.

Original languageEnglish (US)
Title of host publication2022 IEEE International Conference on Robotics and Automation, ICRA 2022
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages4414-4421
Number of pages8
ISBN (Electronic)9781728196817
DOIs
StatePublished - 2022
Externally publishedYes
Event39th IEEE International Conference on Robotics and Automation, ICRA 2022 - Philadelphia, United States
Duration: May 23 2022May 27 2022

Publication series

NameProceedings - IEEE International Conference on Robotics and Automation
ISSN (Print)1050-4729

Conference

Conference39th IEEE International Conference on Robotics and Automation, ICRA 2022
Country/TerritoryUnited States
CityPhiladelphia
Period5/23/225/27/22

All Science Journal Classification (ASJC) codes

  • Software
  • Artificial Intelligence
  • Electrical and Electronic Engineering
  • Control and Systems Engineering

Fingerprint

Dive into the research topics of 'Stronger Generalization Guarantees for Robot Learning by Combining Generative Models and Real-World Data'. Together they form a unique fingerprint.

Cite this