Improving Foundation Models for Few-Shot Learning via Multitask Finetuning

Abstract

Foundation models have become essential tools for AI. In this paper, we study the problem of adapting foundation models, pre-trained using contrastive learning, to downstream tasks with limited labels. We explore the paradigm of finetuning a foundation model before adapting to a target task, using a set of related tasks with a few labeled samples. We show both theoretically and empirically that with a diverse set of related tasks this finetuning leads to reduced error in the target task, when compared with directly adapting the same pre-trained model, e.g., at least 6\% target accuracy improvements on the miniImageNet.

Cite

Text

Xu et al. "Improving Foundation Models for Few-Shot Learning via Multitask Finetuning." ICLR 2023 Workshops: ME-FoMo, 2023.

Markdown

[Xu et al. "Improving Foundation Models for Few-Shot Learning via Multitask Finetuning." ICLR 2023 Workshops: ME-FoMo, 2023.](https://mlanthology.org/iclrw/2023/xu2023iclrw-improving/)

BibTeX

@inproceedings{xu2023iclrw-improving,
  title     = {{Improving Foundation Models for Few-Shot Learning via Multitask Finetuning}},
  author    = {Xu, Zhuoyan and Shi, Zhenmei and Wei, Junyi and Li, Yin and Liang, Yingyu},
  booktitle = {ICLR 2023 Workshops: ME-FoMo},
  year      = {2023},
  url       = {https://mlanthology.org/iclrw/2023/xu2023iclrw-improving/}
}