# Supervised and Unsupervised Neural Approaches to Text Readability

## TL;DR

This paper explores neural supervised and unsupervised methods for assessing document readability, demonstrating robustness, transferability, and comprehensive analysis across multiple languages and datasets.

## Contribution

It introduces novel neural approaches for readability assessment and provides a systematic comparison of architectures and their performance against existing methods.

## Key findings

- Unsupervised neural approach is robust and transferable across languages.
- Neural architectures outperform traditional feature-based methods in many cases.
- The study offers insights into strengths and weaknesses of different neural models.

## Abstract

We present a set of novel neural supervised and unsupervised approaches for determining the readability of documents. In the unsupervised setting, we leverage neural language models, whereas in the supervised setting, three different neural classification architectures are tested. We show that the proposed neural unsupervised approach is robust, transferable across languages and allows adaptation to a specific readability task and data set. By systematic comparison of several neural architectures on a number of benchmark and new labelled readability datasets in two languages, this study also offers a comprehensive analysis of different neural approaches to readability classification. We expose their strengths and weaknesses, compare their performance to current state-of-the-art classification approaches to readability, which in most cases still rely on extensive feature engineering, and propose possibilities for improvements.