Investigating Text Simplification Evaluation
This work addresses evaluation challenges in text simplification for NLP researchers, though it appears incremental as it builds on existing analysis methods.
The paper investigated text simplification evaluation by analyzing parallel corpora and found they contain inaccurate simplifications, incorrect alignments, and significant differences between training and test subsets. The authors demonstrated that improving dataset distribution leads to more robust text simplification models.
Modern text simplification (TS) heavily relies on the availability of gold standard data to build machine learning models. However, existing studies show that parallel TS corpora contain inaccurate simplifications and incorrect alignments. Additionally, evaluation is usually performed by using metrics such as BLEU or SARI to compare system output to the gold standard. A major limitation is that these metrics do not match human judgements and the performance on different datasets and linguistic phenomena vary greatly. Furthermore, our research shows that the test and training subsets of parallel datasets differ significantly. In this work, we investigate existing TS corpora, providing new insights that will motivate the improvement of existing state-of-the-art TS evaluation methods. Our contributions include the analysis of TS corpora based on existing modifications used for simplification and an empirical study on TS models performance by using better-distributed datasets. We demonstrate that by improving the distribution of TS datasets, we can build more robust TS models.