Pre-trained masked language models successfully perform few-shot learning by formulating downstream tasks as text infilling. However, as a strong alternative in full-shot settings, discriminative pre-trained models like ELECTRA do not fit into the paradigm. In this work, we adapt prompt-based few-shot learning to ELECTRA and show that it outperforms masked language models in a wide range of tasks. ELECTRA is pre-trained to distinguish if a token is generated or original. We naturally extend that to prompt-based few-shot learning by training to score the originality of the target options without introducing new parameters. Our method can be easily adapted to tasks involving multi-token predictions without extra computation overhead. Analysis shows that ELECTRA learns distributions that align better with downstream tasks.
@article{arxiv.2205.15223,
title = {Prompting ELECTRA: Few-Shot Learning with Discriminative Pre-Trained Models},
author = {Mengzhou Xia and Mikel Artetxe and Jingfei Du and Danqi Chen and Ves Stoyanov},
journal= {arXiv preprint arXiv:2205.15223},
year = {2022}
}
Comments
Accepted to EMNLP 2022; The code is available at https://github.com/facebookresearch/ELECTRA-Fewshot-Learning