Dynamic Planning Networks
This addresses the challenge of costly trial-and-error in reinforcement learning for agents, though it appears incremental as it builds on existing model-based and model-free approaches.
The paper tackles the problem of inefficient planning in deep reinforcement learning by introducing Dynamic Planning Networks (DPN), which combine model-based and model-free aspects to dynamically construct plans, reducing required state-transitions by up to 96% and improving data efficiency, performance, and generalization.
We introduce Dynamic Planning Networks (DPN), a novel architecture for deep reinforcement learning, that combines model-based and model-free aspects for online planning. Our architecture learns to dynamically construct plans using a learned state-transition model by selecting and traversing between simulated states and actions to maximize information before acting. In contrast to model-free methods, model-based planning lets the agent efficiently test action hypotheses without performing costly trial-and-error in the environment. DPN learns to efficiently form plans by expanding a single action-conditional state transition at a time instead of exhaustively evaluating each action, reducing the required number of state-transitions during planning by up to 96%. We observe various emergent planning patterns used to solve environments, including classical search methods such as breadth-first and depth-first search. DPN shows improved data efficiency, performance, and generalization to new and unseen domains in comparison to several baselines.