Improving the Sample Efficiency of Prompt Tuning with Domain Adaptation

TL;DR

This paper introduces OPTIMA, a domain adaptation method that improves prompt tuning's transferability and sample efficiency by regularizing decision boundaries, especially effective in few-shot and data-scarce scenarios.

Contribution

The paper proposes OPTIMA, a novel domain adaptation technique for prompt tuning that leverages unlabeled target domain data to enhance transferability and efficiency.

Findings

OPTIMA outperforms strong baselines in transfer tasks.

OPTIMA significantly improves sample efficiency in prompt tuning.

In few-shot settings, OPTIMA surpasses full-model tuning.

Abstract

Prompt tuning, or the conditioning of a frozen pretrained language model (PLM) with soft prompts learned from data, has demonstrated impressive performance on a wide range of NLP tasks. However, prompt tuning requires a large training dataset to be effective and is outperformed by finetuning the entire PLM in data-scarce regimes. Previous work (Gu et al., 2022, Vu et al., 2022) proposed to transfer soft prompts pretrained on the source domain to the target domain. In this paper, we explore domain adaptation for prompt tuning, a problem setting where unlabeled data from the target domain are available during pretraining. We propose bOosting Prompt TunIng with doMain Adaptation (OPTIMA), which regularizes the decision boundary to be smooth around regions where source and target data distributions are similar. Extensive experiments demonstrate that OPTIMA significantly enhances the transferability and sample-efficiency of prompt tuning compared to strong baselines. Moreover, in few-shot settings, OPTIMA exceeds full-model tuning by a large margin.