batra-mlp-lab / va_mbrl

Value-aware model-based reinforcement learning

Model-Advantage and Value-Aware Models for Model-Based Reinforcement Learning: Bridging the Gap in Theory and Practice

This repository contains the code for the paper Model-Advantage and Value-Aware Models for Model-Based Reinforcement Learning: Bridging the Gap in Theory and Practice. This repository has two submodules that correspond to the code for the SLBO and MBPO experiments in the paper.

Part A: VA-SLBO Repository

Part B: VA-MBPO Repository