Emergent Neural Automaton Policies: Learning Symbolic Structure from Visuomotor Trajectories

TL;DR

ENAP introduces an adaptive neuro-symbolic framework that learns interpretable task structures from visuomotor data, improving sample efficiency and performance in long-horizon robotic tasks.

Contribution

It combines adaptive clustering and automaton inference to derive a high-level symbolic planner guiding low-level control, a novel approach in robot learning.

Findings

ENAP outperforms state-of-the-art end-to-end policies by up to 27% in low-data regimes.

It achieves high sample efficiency and interpretability without task-specific labels.

The framework effectively captures latent task modes through learned automata.

Abstract

Scaling robot learning to long-horizon tasks remains a formidable challenge. While end-to-end policies often lack the structural priors needed for effective long-term reasoning, traditional neuro-symbolic methods rely heavily on hand-crafted symbolic priors. To address the issue, we introduce ENAP (Emergent Neural Automaton Policy), a framework that allows a bi-level neuro-symbolic policy adaptively emerge from visuomotor demonstrations. Specifically, we first employ adaptive clustering and an extension of the L* algorithm to infer a Mealy state machine from visuomotor data, which serves as an interpretable high-level planner capturing latent task modes. Then, this discrete structure guides a low-level reactive residual network to learn precise continuous control via behavior cloning (BC). By explicitly modeling the task structure with discrete transitions and continuous residuals, ENAP achieves high sample efficiency and interpretability without requiring task-specific labels. Extensive experiments on complex manipulation and long-horizon tasks demonstrate that ENAP outperforms state-of-the-art (SoTA) end-to-end VLA policies by up to 27% in low-data regimes, while offering a structured representation of robotic intent (Fig. 1).