Forward-Backward Reinforcement Learning

TL;DR

This paper introduces a backward reinforcement learning approach that leverages goal states to improve training efficiency by predicting reverse steps, demonstrated to outperform standard methods in Gridworld and Towers of Hanoi.

Contribution

The paper proposes a novel backward reinforcement learning method that uses imagined reversal steps from goal states to accelerate training.

Findings

Better performance than standard DDQN in experiments

Effective in Gridworld and Towers of Hanoi environments

Leverages goal states to improve learning efficiency

Abstract

Goals for reinforcement learning problems are typically defined through hand-specified rewards. To design such problems, developers of learning algorithms must inherently be aware of what the task goals are, yet we often require agents to discover them on their own without any supervision beyond these sparse rewards. While much of the power of reinforcement learning derives from the concept that agents can learn with little guidance, this requirement greatly burdens the training process. If we relax this one restriction and endow the agent with knowledge of the reward function, and in particular of the goal, we can leverage backwards induction to accelerate training. To achieve this, we propose training a model to learn to take imagined reversal steps from known goal states. Rather than training an agent exclusively to determine how to reach a goal while moving forwards in time, our approach travels backwards to jointly predict how we got there. We evaluate our work in Gridworld and Towers of Hanoi and empirically demonstrate that it yields better performance than standard DDQN.