How May I Help You? Using Neural Text Simplification to Improve Downstream NLP Tasks

TL;DR

This paper explores neural text simplification as a tool to enhance NLP tasks by either simplifying input texts or augmenting training data, showing improvements in relation extraction and natural language inference.

Contribution

It introduces the novel application of neural text simplification for data augmentation in NLP, demonstrating performance gains on relation extraction and NLI tasks.

Findings

Data augmentation with neural TS improves relation extraction accuracy.

Simplifying input texts at prediction time has limited impact.

Neural TS enhances NLP task performance across multiple datasets.

Abstract

The general goal of text simplification (TS) is to reduce text complexity for human consumption. This paper investigates another potential use of neural TS: assisting machines performing natural language processing (NLP) tasks. We evaluate the use of neural TS in two ways: simplifying input texts at prediction time and augmenting data to provide machines with additional information during training. We demonstrate that the latter scenario provides positive effects on machine performance on two separate datasets. In particular, the latter use of TS improves the performances of LSTM (1.82-1.98%) and SpanBERT (0.7-1.3%) extractors on TACRED, a complex, large-scale, real-world relation extraction task. Further, the same setting yields improvements of up to 0.65% matched and 0.62% mismatched accuracies for a BERT text classifier on MNLI, a practical natural language inference dataset.