Prompt to be Consistent is Better than Self-Consistent? Few-Shot and Zero-Shot Fact Verification with Pre-trained Language Models

TL;DR

ProToCo is a novel method that enhances pre-trained language models' fact verification accuracy in few-shot and zero-shot scenarios by generating claim variants and enforcing consistency constraints, outperforming existing baselines.

Contribution

The paper introduces ProToCo, a new approach that improves fact verification in low-resource settings through consistency-based prompt tuning and parameter-efficient fine-tuning.

Findings

ProToCo outperforms state-of-the-art few-shot fact verification methods.

ProToCo surpasses strong zero-shot learners like T0 with minimal unlabeled data.

ProToCo exceeds large PLMs using in-context learning in both few- and zero-shot settings.

Abstract

Few-shot or zero-shot fact verification only relies on a few or no labeled training examples. In this paper, we propose a novel method called ProToCo, to \underline{Pro}mpt pre-trained language models (PLMs) \underline{To} be \underline{Co}nsistent, for improving the factuality assessment capability of PLMs in the few-shot and zero-shot settings. Given a claim-evidence pair, ProToCo generates multiple variants of the claim with different relations and frames a simple consistency mechanism as constraints for making compatible predictions across these variants. We update PLMs by using parameter-efficient fine-tuning (PEFT), leading to more accurate predictions in few-shot and zero-shot fact verification tasks. Our experiments on three public verification datasets show that ProToCo significantly outperforms state-of-the-art few-shot fact verification baselines. With a small number of unlabeled instances, ProToCo also outperforms the strong zero-shot learner T0 on zero-shot verification. Compared to large PLMs using in-context learning (ICL) method, ProToCo outperforms OPT-30B and the Self-Consistency-enabled OPT-6.7B model in both few- and zero-shot settings.