An LLM-Enhanced Adversarial Editing System for Lexical Simplification

TL;DR

This paper introduces a novel lexical simplification system that leverages adversarial editing and large language models to simplify text without relying on parallel corpora, showing strong results on benchmark datasets.

Contribution

It presents a new LS approach combining adversarial editing with LLM knowledge distillation, effective in low-resource settings and without needing parallel data.

Findings

Outperforms existing LS methods on benchmark datasets

Effective in low-resource scenarios without parallel corpora

Demonstrates the benefit of LLM guidance in lexical simplification

Abstract

Lexical Simplification (LS) aims to simplify text at the lexical level. Existing methods rely heavily on annotated data, making it challenging to apply in low-resource scenarios. In this paper, we propose a novel LS method without parallel corpora. This method employs an Adversarial Editing System with guidance from a confusion loss and an invariance loss to predict lexical edits in the original sentences. Meanwhile, we introduce an innovative LLM-enhanced loss to enable the distillation of knowledge from Large Language Models (LLMs) into a small-size LS system. From that, complex words within sentences are masked and a Difficulty-aware Filling module is crafted to replace masked positions with simpler words. At last, extensive experimental results and analyses on three benchmark LS datasets demonstrate the effectiveness of our proposed method.