Semantically Driven Sentence Fusion: Modeling and Evaluation

TL;DR

This paper introduces a semantically driven sentence fusion approach that uses multiple reference solutions and auxiliary tasks to improve model robustness and semantic understanding in sentence fusion tasks.

Contribution

It proposes automatically expanding ground-truths into multiple references and incorporating auxiliary discourse classification to enhance sentence fusion models.

Findings

Improved model performance over state-of-the-art methods

Enhanced semantic capturing through multiple references

Effective multi-task learning framework for sentence fusion

Abstract

Sentence fusion is the task of joining related sentences into coherent text. Current training and evaluation schemes for this task are based on single reference ground-truths and do not account for valid fusion variants. We show that this hinders models from robustly capturing the semantic relationship between input sentences. To alleviate this, we present an approach in which ground-truth solutions are automatically expanded into multiple references via curated equivalence classes of connective phrases. We apply this method to a large-scale dataset and use the augmented dataset for both model training and evaluation. To improve the learning of semantic representation using multiple references, we enrich the model with auxiliary discourse classification tasks under a multi-tasking framework. Our experiments highlight the improvements of our approach over state-of-the-art models.