SINDy-RL: Interpretable and Efficient Model-Based Reinforcement Learning
This addresses the need for more efficient and interpretable reinforcement learning in applications like flow control, though it appears incremental as it builds on existing SINDy and DRL methods.
The paper tackles the problem of deep reinforcement learning (DRL) being data-intensive and producing uninterpretable policies by introducing SINDy-RL, a framework that combines sparse identification of nonlinear dynamics (SINDy) with DRL to create efficient and interpretable models. It achieves comparable performance to DRL on benchmark control environments and flow control problems, using significantly fewer environment interactions and producing policies orders of magnitude smaller.
Deep reinforcement learning (DRL) has shown significant promise for uncovering sophisticated control policies that interact in complex environments, such as stabilizing a tokamak fusion reactor or minimizing the drag force on an object in a fluid flow. However, DRL requires an abundance of training examples and may become prohibitively expensive for many applications. In addition, the reliance on deep neural networks often results in an uninterpretable, black-box policy that may be too computationally expensive to use with certain embedded systems. Recent advances in sparse dictionary learning, such as the sparse identification of nonlinear dynamics (SINDy), have shown promise for creating efficient and interpretable data-driven models in the low-data regime. In this work we introduce SINDy-RL, a unifying framework for combining SINDy and DRL to create efficient, interpretable, and trustworthy representations of the dynamics model, reward function, and control policy. We demonstrate the effectiveness of our approaches on benchmark control environments and flow control problems, including gust mitigation on a 3D NACA 0012 airfoil at $Re=1000$. SINDy-RL achieves comparable performance to modern DRL algorithms using significantly fewer interactions in the environment and results in an interpretable control policy orders of magnitude smaller than a DRL policy.