Dynamic Action Interpolation: A Universal Approach for Accelerating Reinforcement Learning with Expert Guidance
This addresses the sample inefficiency issue in reinforcement learning for researchers and practitioners, offering a universal and straightforward solution without complex modifications.
The paper tackles the sample inefficiency problem in reinforcement learning by proposing Dynamic Action Interpolation (DAI), a simple framework that integrates expert guidance into Actor-Critic algorithms, resulting in average early-stage performance improvements of over 160% and final performance gains of more than 50% across MuJoCo tasks.
Reinforcement learning (RL) suffers from severe sample inefficiency, especially during early training, requiring extensive environmental interactions to perform competently. Existing methods tend to solve this by incorporating prior knowledge, but introduce significant architectural and implementation complexity. We propose Dynamic Action Interpolation (DAI), a universal yet straightforward framework that interpolates expert and RL actions via a time-varying weight $α(t)$, integrating into any Actor-Critic algorithm with just a few lines of code and without auxiliary networks or additional losses. Our theoretical analysis shows that DAI reshapes state visitation distributions to accelerate value function learning while preserving convergence guarantees. Empirical evaluations across MuJoCo continuous control tasks demonstrate that DAI improves early-stage performance by over 160\% on average and final performance by more than 50\%, with the Humanoid task showing a 4$\times$ improvement early on and a 2$\times$ gain at convergence. These results challenge the assumption that complex architectural modifications are necessary for sample-efficient reinforcement learning.