---
id: 20261118-T0-11
title: "参数合并让VLA模型找回语言能力：机器人指令跟随不再退化"
title_en: "Parameter Merging Restores VLM Language Skills in VLA Models"
url: https://ai.daily.yangsir.net/daily/20261118-T0-11
issue_date: 2026-11-18
publish_date: 2026-11-17T05:00:00.000Z
category: research
source_name: "arXiv cs.AI"
source_url: https://arxiv.org/abs/2611.18891
---

# 参数合并让VLA模型找回语言能力：机器人指令跟随不再退化

arXiv新论文研究视觉-语言-动作模型（VLA）的一个退化问题：VLA由预训练视觉语言模型（VLM）在机器人示范数据上微调而来，但微调会导致其语言能力退化。研究提出通过参数合并把语言能力找回来，在保留动作预测能力的同时恢复VLM原有的语言理解。这有助于机器人在执行动作时更好地理解语言指令。

## English Version

**Parameter Merging Restores VLM Language Skills in VLA Models**

A new arXiv paper examines a degradation problem in Vision-Language-Action models (VLAs): VLAs are fine-tuned from pretrained Vision-Language Models (VLMs) on robot demonstration data, and that fine-tuning erodes language ability. The work uses parameter merging to restore language capabilities while keeping action prediction intact. Better language grounding helps robots follow instructions while executing actions.

---

**来源**：[arXiv cs.AI](https://arxiv.org/abs/2611.18891)

**详情页**：https://ai.daily.yangsir.net/daily/20261118-T0-11

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*