PPT: Pre-trained Prompt Tuning for Few-shot Learning

TL;DR

This paper introduces PPT, a pre-trained prompt tuning framework that improves few-shot learning performance of large-scale language models by pre-training soft prompts, outperforming traditional fine-tuning methods.

Contribution

The paper proposes pre-training soft prompts before task-specific tuning, enhancing prompt tuning's effectiveness in few-shot scenarios compared to existing methods.

Findings

Pre-trained prompts outperform fine-tuning in few-shot settings.

PPT achieves comparable or better results than full-model fine-tuning.

The approach is efficient and generalizes well across tasks.

Abstract

Prompts for pre-trained language models (PLMs) have shown remarkable performance by bridging the gap between pre-training tasks and various downstream tasks. Among these methods, prompt tuning, which freezes PLMs and only tunes soft prompts, provides an efficient and effective solution for adapting large-scale PLMs to downstream tasks. However, prompt tuning is yet to be fully explored. In our pilot experiments, we find that prompt tuning performs comparably with conventional full-model fine-tuning when downstream data are sufficient, whereas it performs much worse under few-shot learning settings, which may hinder the application of prompt tuning in practice. We attribute this low performance to the manner of initializing soft prompts. Therefore, in this work, we propose to pre-train prompts by adding soft prompts into the pre-training stage to obtain a better initialization. We name this Pre-trained Prompt Tuning framework "PPT". To ensure the generalization of PPT, we formulate similar classification tasks into a unified task form and pre-train soft prompts for this unified task. Extensive experiments show that tuning pre-trained prompts for downstream tasks can reach or even outperform full-model fine-tuning under both full-data and few-shot settings. Our approach is effective and efficient for using large-scale PLMs in practice.