English

Robust Regression for Safe Exploration in Control

Machine Learning 2020-06-30 v2 Systems and Control Systems and Control Machine Learning

Abstract

We study the problem of safe learning and exploration in sequential control problems. The goal is to safely collect data samples from operating in an environment, in order to learn to achieve a challenging control goal (e.g., an agile maneuver close to a boundary). A central challenge in this setting is how to quantify uncertainty in order to choose provably-safe actions that allow us to collect informative data and reduce uncertainty, thereby achieving both improved controller safety and optimality. To address this challenge, we present a deep robust regression model that is trained to directly predict the uncertainty bounds for safe exploration. We derive generalization bounds for learning, and connect them with safety and stability bounds in control. We demonstrate empirically that our robust regression approach can outperform the conventional Gaussian process (GP) based safe exploration in settings where it is difficult to specify a good GP prior.

Keywords

Cite

@article{arxiv.1906.05819,
  title  = {Robust Regression for Safe Exploration in Control},
  author = {Anqi Liu and Guanya Shi and Soon-Jo Chung and Anima Anandkumar and Yisong Yue},
  journal= {arXiv preprint arXiv:1906.05819},
  year   = {2020}
}

Comments

2nd Annual Conference on Learning for Dynamics and Control

R2 v1 2026-06-23T09:53:03.239Z