Probing Factually Grounded Content Transfer with Factual Ablation

TL;DR

This paper introduces a new automatic evaluation method for factual consistency in grounded text generation, using factual ablation to measure how relevance of grounding affects output, and proposes improvements over existing baselines.

Contribution

The paper presents the concept of factual ablation for automatic factuality measurement and introduces new evaluation sets and methods for content transfer tasks.

Findings

Factual ablation effectively measures factual consistency.

New evaluation sets enable better assessment of grounded generation.

Proposed methods outperform strong baselines.

Abstract

Despite recent success, large neural models often generate factually incorrect text. Compounding this is the lack of a standard automatic evaluation for factuality--it cannot be meaningfully improved if it cannot be measured. Grounded generation promises a path to solving both of these problems: models draw on a reliable external document (grounding) for factual information, simplifying the challenge of factuality. Measuring factuality is also simplified--to factual consistency, testing whether the generation agrees with the grounding, rather than all facts. Yet, without a standard automatic metric for factual consistency, factually grounded generation remains an open problem. We study this problem for content transfer, in which generations extend a prompt, using information from factual grounding. Particularly, this domain allows us to introduce the notion of factual ablation for automatically measuring factual consistency: this captures the intuition that the model should be less likely to produce an output given a less relevant grounding document. In practice, we measure this by presenting a model with two grounding documents, and the model should prefer to use the more factually relevant one. We contribute two evaluation sets to measure this. Applying our new evaluation, we propose multiple novel methods improving over strong baselines.