A Neural Pairwise Ranking Model for Readability Assessment

TL;DR

This paper introduces the first neural pairwise ranking model for automatic readability assessment, demonstrating strong monolingual and cross-lingual performance, and releasing a new bilingual dataset.

Contribution

It presents a novel neural pairwise ranking approach for ARA and provides the first cross-lingual zero-shot evaluation results.

Findings

Achieves over 80% cross-lingual ranking accuracy in zero-shot settings.

Performs well in monolingual and cross-corpus testing scenarios.

Introduces a new English-French bilingual readability dataset.

Abstract

Automatic Readability Assessment (ARA), the task of assigning a reading level to a text, is traditionally treated as a classification problem in NLP research. In this paper, we propose the first neural, pairwise ranking approach to ARA and compare it with existing classification, regression, and (non-neural) ranking methods. We establish the performance of our model by conducting experiments with three English, one French and one Spanish datasets. We demonstrate that our approach performs well in monolingual single/cross corpus testing scenarios and achieves a zero-shot cross-lingual ranking accuracy of over 80% for both French and Spanish when trained on English data. Additionally, we also release a new parallel bilingual readability dataset in English and French. To our knowledge, this paper proposes the first neural pairwise ranking model for ARA, and shows the first results of cross-lingual, zero-shot evaluation of ARA with neural models.