Differentiate-and-Inject: Enhancing VLAs via Functional Differentiation Induced by In-Parameter Structural Reasoning

作者：Jingyi Hou, Leyu Zhou, Chenchen Jing, Jinghan Yang, Xinbo Yu, Wei He · 单位：University of Science and Technology Beijing Zhejiang Uni · 会议/期刊：arXiv · 日期：2026-02-07 · 来源：Low-Level Learning-Based Action Modelling / Input Modelling / 2D Vision Language Action Models with Auxiliary Tasks - Visual Goal Extraction