Contrastive Learning for Prompt-Based Few-Shot Language Learners

TL;DR

This paper introduces a supervised contrastive learning framework for prompt-based few-shot language models, improving their generalization by clustering same-class inputs with different prompts and demonstrating superior results across multiple NLP tasks.

Contribution

It proposes a novel contrastive learning approach integrated with prompt-based fine-tuning, enhancing few-shot learning performance with minimal assumptions.

Findings

Outperforms state-of-the-art methods on 15 NLP tasks

Effectively clusters same-class inputs with different prompts

Can be integrated with various prompt-based models

Abstract

The impressive performance of GPT-3 using natural language prompts and in-context learning has inspired work on better fine-tuning of moderately-sized models under this paradigm. Following this line of work, we present a contrastive learning framework that clusters inputs from the same class for better generality of models trained with only limited examples. Specifically, we propose a supervised contrastive framework that clusters inputs from the same class under different augmented "views" and repel the ones from different classes. We create different "views" of an example by appending it with different language prompts and contextual demonstrations. Combining a contrastive loss with the standard masked language modeling (MLM) loss in prompt-based few-shot learners, the experimental results show that our method can improve over the state-of-the-art methods in a diverse set of 15 language tasks. Our framework makes minimal assumptions on the task or the base model, and can be applied to many recent methods with little modification. The code will be made available at: https://github.com/yiren-jian/LM-SupCon.