PropSegmEnt: A Large-Scale Corpus for Proposition-Level Segmentation and Entailment Recognition

TL;DR

PropSegmEnt introduces a large annotated corpus for proposition-level segmentation and entailment recognition, enabling more granular analysis of entailment relations within sentences for improved natural language inference understanding.

Contribution

The paper presents a novel dataset with proposition-level annotations and establishes baseline models for segmentation and entailment tasks, advancing fine-grained NLI research.

Findings

Strong baseline performance on segmentation and entailment tasks

Potential for improved NLI explainability and compositionality analysis

Usefulness demonstrated in summary hallucination detection

Abstract

The widely studied task of Natural Language Inference (NLI) requires a system to recognize whether one piece of text is textually entailed by another, i.e. whether the entirety of its meaning can be inferred from the other. In current NLI datasets and models, textual entailment relations are typically defined on the sentence- or paragraph-level. However, even a simple sentence often contains multiple propositions, i.e. distinct units of meaning conveyed by the sentence. As these propositions can carry different truth values in the context of a given premise, we argue for the need to recognize the textual entailment relation of each proposition in a sentence individually. We propose PropSegmEnt, a corpus of over 45K propositions annotated by expert human raters. Our dataset structure resembles the tasks of (1) segmenting sentences within a document to the set of propositions, and (2) classifying the entailment relation of each proposition with respect to a different yet topically-aligned document, i.e. documents describing the same event or entity. We establish strong baselines for the segmentation and entailment tasks. Through case studies on summary hallucination detection and document-level NLI, we demonstrate that our conceptual framework is potentially useful for understanding and explaining the compositionality of NLI labels.