Contrastive Learning for Prompt-Based Few-Shot Language Learners
This work addresses the challenge of enhancing model generality with limited data for researchers and practitioners in natural language processing, representing an incremental improvement over existing prompt-based methods.
The paper tackles the problem of improving few-shot learning for language models using prompts by introducing a contrastive learning framework that clusters inputs from the same class under different augmented views, resulting in state-of-the-art performance across 15 diverse language tasks.
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.