NS-VLA: Towards Neuro-Symbolic Vision-Language-Action Models
This work addresses data efficiency and exploration limitations in robotic manipulation for researchers and practitioners, though it appears incremental by combining neuro-symbolic methods with reinforcement learning.
The paper tackles challenges in Vision-Language-Action models for robotic manipulation by proposing a neuro-symbolic framework that outperforms previous methods in one-shot training and data-perturbed settings, achieving superior zero-shot generalizability, high data efficiency, and expanded exploration space.
Vision-Language-Action (VLA) models are formulated to ground instructions in visual context and generate action sequences for robotic manipulation. Despite recent progress, VLA models still face challenges in learning related and reusable primitives, reducing reliance on large-scale data and complex architectures, and enabling exploration beyond demonstrations. To address these challenges, we propose a novel Neuro-Symbolic Vision-Language-Action (NS-VLA) framework via online reinforcement learning (RL). It introduces a symbolic encoder to embedding vision and language features and extract structured primitives, utilizes a symbolic solver for data-efficient action sequencing, and leverages online RL to optimize generation via expansive exploration. Experiments on robotic manipulation benchmarks demonstrate that NS-VLA outperforms previous methods in both one-shot training and data-perturbed settings, while simultaneously exhibiting superior zero-shot generalizability, high data efficiency and expanded exploration space. Our code is available.