Forward-Backward Reinforcement Learning
arXiv:1803.10227
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.
References in corpus (3)
Cited by in corpus (11)
- RUDDER: Return Decomposition for Delayed Rewards
- Imitating Latent Policies from Observation
- Recall Traces: Backtracking Models for Efficient Reinforcement Learning
- Backplay: "Man muss immer umkehren"
- Offline Reinforcement Learning with Reverse Model-based Imagination
- Learning Compositional Neural Programs with Recursive Tree Search and Planning
- PlayVirtual: Augmenting Cycle-Consistent Virtual Trajectories for Reinforcement Learning
- Estimating Q(s,s') with Deep Deterministic Dynamics Gradients
- Competitive Experience Replay
- On the Parallel Tower of Hanoi Puzzle: Acyclicity and a Conditional Triangle Inequality
- Solving Sokoban with forward-backward reinforcement learning