Sample-efficient diffusion-based control of complex nonlinear systems

ICLR 2026 Conference Submission16678 Authors

Keywords: Complex System, Data-driven Control, Generative Model, Diffusion Model, AI for Science

TL;DR: We purpose a new diffusion-based controller that masters complex nonlinear systems with 90% less training data than state-of-the-art methods.

Abstract: Complex nonlinear system control faces challenges in achieving sample-efficient, reliable performance. While diffusion-based methods have demonstrated advantages over classical and reinforcement learning approaches in long-term control performance, they are limited by sample efficiency. This paper presents SEDC (Sample-Efficient Diffusion-based Control), a novel diffusion-based control framework addressing three core challenges: high-dimensional state-action spaces, nonlinear system dynamics, and the gap between non-optimal training data and near-optimal control solutions. Our approach introduces a novel control paradigm by architecturally decoupling state-action learning and decomposing dynamics, while a guided self-finetuning process iteratively refines the control policy. These coordinated innovations allow SEDC to achieve 39.5\%-47.3\% better control accuracy than baselines while using only 10\% of the training samples, as validated across multiple complex nonlinear dynamic systems. Our approach represents a significant advancement in sample-efficient control of complex nonlinear systems. The implementation of the code can be found \href{https://anonymous.4open.science/r/DIFOCON-C019}{here}.

Primary Area: applications to physical sciences (physics, chemistry, biology, etc.)

Submission Number: 16678