充分利用现有资源：在低数据条件下调整预训练的视觉语言模型（英文版）

英文标题：Making the Most of What You Have: Adapting Pre-trained Visual Language Models in the Low-data Regime中文摘要：研究表明，视觉语言模型是目前广泛使用的预训练模型，但在适应少量样本方面，深度学习模型存在不足。本文研究了面向生成视觉语言模型的现有适应方法，提出了自标记的重要性，并提出一种任务适应流水线，可显著提高各种视觉语言任务（如图像分类、视觉问答等）的性能。英文摘要：Large-scale visual language models are widely used as pre-trained models andthen adapted for various downstream tasks. While humans are known toefficiently learn new tasks from a few examples, deep learning models strugglewith adaptation from few examples. In