Entailment as Few-Shot Learner

29 Apr 2021  ·  Sinong Wang, Han Fang, Madian Khabsa, Hanzi Mao, Hao Ma ·

Large pre-trained language models (LMs) have demonstrated remarkable ability as few-shot learners. However, their success hinges largely on scaling model parameters to a degree that makes it challenging to train and serve. In this paper, we propose a new approach, named as EFL, that can turn small LMs into better few-shot learners. The key idea of this approach is to reformulate potential NLP task into an entailment one, and then fine-tune the model with as little as 8 examples. We further demonstrate our proposed method can be: (i) naturally combined with an unsupervised contrastive learning-based data augmentation method; (ii) easily extended to multilingual few-shot learning. A systematic evaluation on 18 standard NLP tasks demonstrates that this approach improves the various existing SOTA few-shot learning methods by 12\%, and yields competitive few-shot performance with 500 times larger models, such as GPT-3.

PDF Abstract

Results from the Paper

Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Question Answering BoolQ EFL Accuracy 86.0 # 6
Linguistic Acceptability CoLA EFL Accuracy 86.4% # 1
Sentiment Analysis CR EFL Accuracy 92.5 # 2
Sentiment Analysis IMDb EFL Accuracy 96.1 # 3
Sentiment Analysis MPQA EFL Accuracy 90.8 # 1
Sentiment Analysis MR EFL Accuracy 92.5 # 2
Semantic Textual Similarity MRPC EFL F1 91.0 # 7
Topic Classification OS EFL Accuracy 95.1 # 1
Natural Language Inference QNLI EFL Accuracy 94.5% # 14
Paraphrase Identification Quora Question Pairs EFL F1 89.2 # 1
Natural Language Inference RTE EFL Accuracy 90.5% # 7
Natural Language Inference SNLI EFL % Test Accuracy 93.1 # 1
Sentiment Analysis SST-2 Binary classification EFL Accuracy 96.9 # 8
Semantic Textual Similarity STS Benchmark EFL Pearson Correlation 0.918 # 10
Subjectivity Analysis SUBJ EFL Accuracy 97.1 # 3