Timezone: »

 
Poster
Neurosymbolic Reinforcement Learning with Formally Verified Exploration
Greg Anderson · Abhinav Verma · Isil Dillig · Swarat Chaudhuri

Tue Dec 08 09:00 AM -- 11:00 AM (PST) @ Poster Session 1 #202
in Poster Session 2 »

We present REVEL, a partially neural reinforcement learning (RL) framework for provably safe exploration in continuous state and action spaces. A key challenge for provably safe deep RL is that repeatedly verifying neural networks within a learning loop is computationally infeasible. We address this challenge using two policy classes: a general, neurosymbolic class with approximate gradients and a more restricted class of symbolic policies that allows efficient verification. Our learning algorithm is a mirror descent over policies: in each iteration, it safely lifts a symbolic policy into the neurosymbolic space, performs safe gradient updates to the resulting policy, and projects the updated policy into the safe symbolic subset, all without requiring explicit verification of neural networks. Our empirical results show that REVEL enforces safe exploration in many scenarios in which Constrained Policy Optimization does not, and that it can discover policies that outperform those learned through prior approaches to verified exploration.

Author Information

Greg Anderson (University of Texas at Austin)
Abhinav Verma (University of Texas at Austin)
Isil Dillig (UT Austin)
Swarat Chaudhuri (The University of Texas at Austin)

More from the Same Authors