Model ensemble instead of prompt fusion: a sample-specific knowledge transfer method for few-shot prompt tuning

TL;DR

This paper introduces SESoM, a sample-specific ensemble method for prompt tuning that improves few-shot learning performance by adaptively combining source model predictions, outperforming existing prompt fusion techniques.

Contribution

Proposes SESoM, a novel sample-specific ensemble approach for prompt transfer, enhancing few-shot prompt tuning performance over traditional prompt fusion methods.

Findings

SESoM outperforms existing prompt fusion methods in few-shot tasks.

Ensemble of source models improves generalization in low-data regimes.

Method is effective across multiple NLP tasks and model scales.

Abstract

Prompt tuning approaches, which learn task-specific soft prompts for a downstream task conditioning on frozen pre-trained models, have attracted growing interest due to its parameter efficiency. With large language models and sufficient training data, prompt tuning performs comparably to full-model tuning. However, with limited training samples in few-shot settings, prompt tuning fails to match the performance of full-model fine-tuning. In this work, we focus on improving the few-shot performance of prompt tuning by transferring knowledge from soft prompts of source tasks. Recognizing the good generalization capabilities of ensemble methods in low-data regime, we first experiment and show that a simple ensemble of model predictions based on different source prompts, outperforms existing multi-prompt knowledge transfer approaches such as source prompt fusion in the few-shot setting. Motivated by this observation, we further investigate model ensembles and propose Sample-specific Ensemble of Source Models (SESoM). SESoM learns to adjust the contribution of each source model for each target sample separately when ensembling source model outputs. Through this way, SESoM inherits the superior generalization of model ensemble approaches and simultaneously captures the sample-specific competence of each source prompt. We conduct experiments across a diverse set of eight NLP tasks using models of different scales (T5-{base, large, XL}) and find that SESoM consistently outperforms the existing models of the same as well as larger parametric scale by a large margin.