GIVLA: Deep Geometry Internalization for A Lightweight VLA via Geometry Instruction and Gradient-Informed Training

摘要

Vision-Language-Action (VLA) models often struggle with precise manipulation due to their lack of explicit spatial awareness. To bridge this gap, we propose GIVLA, a framework that internalizes geometric priors into the VLA backbone through a deep-coupled architecture. GIVLA implements a two-stage gradient-informed training paradigm to resolve task-level interference and ensure precise execution, a design grounded in the perception-action duality identified through our analysis of training dynamics. Extensive experiments on LIBERO and physical robot platforms demonstrate that GIVLA achieves superior accuracy and robustness with high parameter efficiency.