Edit-Constrained Decoding for Sentence Simplification

TL;DR

This paper introduces a new edit-constrained decoding method for sentence simplification that enforces stricter lexical constraints, leading to improved simplification quality across multiple datasets.

Contribution

The paper presents a novel edit operation based lexically constrained decoding approach with stricter satisfaction conditions for sentence simplification.

Findings

Outperforms previous methods on three English simplification datasets

Consistently improves simplification quality

Demonstrates effectiveness of stricter lexical constraints

Abstract

We propose edit operation based lexically constrained decoding for sentence simplification. In sentence simplification, lexical paraphrasing is one of the primary procedures for rewriting complex sentences into simpler correspondences. While previous studies have confirmed the efficacy of lexically constrained decoding on this task, their constraints can be loose and may lead to sub-optimal generation. We address this problem by designing constraints that replicate the edit operations conducted in simplification and defining stricter satisfaction conditions. Our experiments indicate that the proposed method consistently outperforms the previous studies on three English simplification corpora commonly used in this task.