In a new preprint, we explored the use of the parametric uncertainty of an agent as a way to learn in risk-sensitive environments. We show improved robustness to changing environment dynamics and lower failure rates in hard locomotion environments.
Here be dragons! 🐉⚠️— Frederik Schubert (@fschubert) June 14, 2021
Automatic Risk Adaptation via Random Network Distillation helps your RL agent to act cautiously in unknown settings. Learn more in our recent preprint: https://t.co/l4WNKxDMPG
A great collaboration with @The_Eimer, B. Rosenhahn and @LindauerMarius.
We will present our work at the Reinforcement Learning for Real Life Workshop @ ICML 2021.