# Recovering Robustness in Model-Free Reinforcement learning

arXiv: https://arxiv.org/abs/1810.09337 · HTML: https://mlsanity.com/paper/1810.09337/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 22 Oct 2018.

## Built on by (most cited first, 5 of 5)

- [How are policy gradient methods affected by the limits of control?](https://mlsanity.com/paper/2206.06863/) (2022)
- [Towards a Theoretical Foundation of Policy Optimization for Learning Control Policies](https://mlsanity.com/paper/2210.04810/) (2022)
- [How Are Learned Perception-Based Controllers Impacted by the Limits of Robust Control?](https://mlsanity.com/paper/2104.00827/) (2021)
- [AKF-SR: Adaptive Kalman Filtering-based Successor Representation](https://mlsanity.com/paper/2204.00049/) (2022)
- [Configuration Path Control](https://mlsanity.com/paper/2204.02471/) (2022)

## Abstract (arXiv)

Reinforcement learning (RL) is used to directly design a control policy using data collected from the system. This paper considers the robustness of controllers trained via model-free RL. The discussion focuses on the standard model-based linear quadratic Gaussian (LQG) problem as a special instance of RL. A simple example, originally formulated for LQG problems, is used to demonstrate that RL with partial observations can lead to poor robustness margins. It is proposed to recover robustness by introducing random perturbations at the system input during the RL training. The perturbation magnitude can be used to trade off performance for robustness. Two simple examples are presented to demonstrate the proposed method for enhancing robustness during RL training.