Anonymous authors
Paper under double-blind review
ABSTRACT
Many practical reinforcement learning problems contain catastrophic states that
the optimal policy visits infrequently or never. Even on toy problems, deep reinforcement
learners periodically revisit these states, once they are forgotten under
a new policy. In this paper, we introduce intrinsic fear, a learned reward shaping
that accelerates deep reinforcement learning and guards oscillating policies against
periodic catastrophes. Our approach incorporates a second model trained via supervised
learning to predict the probability of imminent catastrophe. This score acts
as a penalty on the Q-learning objective. Our theoretical analysis demonstrates that
the perturbed objective yields the same average return under strong assumptions
and an �-close average return under weaker assumptions. Our analysis also shows
robustness to classification errors. Equipped with intrinsic fear, our DQNs solve
the toy environments and improve on the Atari games Seaquest, Asteroids, and
Freeway.