Problem
Research questions and friction points this paper is trying to address.
Goal-Injected Fine-Tuning
Vision-Language-Action models
goal images
computational training cost
model robustness
GIFT: Goal-Injected Fine-Tuning for Efficient Manipulation Policy Adaptation
🤖 AI Summary
本文提出GIFT方法，通过轻量级微调框架将生成的目标图像集成到预训练的VLA模型中，以提高模型在操作任务中的性能和效率。
📝 Abstract
Compared with relying solely on initial observations and language instructions, predicting goal images with generative models as high-level visual guidance can significantly enhance the robustness of Vision-Language-Action (VLA) models. However, most existing foundation models have not systematically incorporated goal image conditioning due to the high computational training cost. To this end, we propose Goal-Injected Fine-Tuning (GIFT), a lightweight and efficient fine-tuning framework that seamlessly integrates generated goal images into multiple representative pretrained VLA models. Our approach introduces goal image features into observations via a zero-initialized convolution which progressively grows parameters from zero and prevents harmful noise from disrupting the pretrained policy during fine-tuning. As training proceeds, goal information is gradually incorporated, enabling efficient goal understanding without disrupting model stability. We further introduce a refined image editing method to generate semantically and visually consistent goal images from initial observations and task instructions. Experiments show that goal-aware VLA models achieve substantial performance gains across tasks: with only a single epoch of fine-tuning, GIFT outperforms the base model by 6.0% and 13.4% on two SIMPLER settings, and by 4.7% on LIBERO, demonstrating both efficiency and effectiveness.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Goal-Injected Fine-Tuning
Zero-Initialized Convolution
Visual-Language-Action Models
Efficient Policy Adaptation
Image Editing Method
🔎 Similar Papers
No similar papers found.
