Prompt-based Learning for Text Readability Assessment

TL;DR

This work demonstrates that pre-trained sequence-to-sequence transformers (T5 and BART) can be adapted to readability assessment by framing it as a pairwise, text-to-text ranking task. By evaluating nine input-output formats and training on parallel datasets (Newsela and OneStopEnglish), the study shows strong cross-domain generalization and high accuracies (up to 99.6% on Newsela and 98.7% on OneStopEnglish) with joint training. The findings highlight the importance of prompt design for readability tasks and suggest that combining diverse readability data can enhance cross-domain performance, albeit with limitations related to pairwise output practicality and sequence length. Overall, the approach points to a flexible, cross-domain methodology for readability assessment using prompt-based learning on seq2seq architectures, with potential extensions to multi-class or regression formulations.

Abstract

We propose the novel adaptation of a pre-trained seq2seq model for readability assessment. We prove that a seq2seq model - T5 or BART - can be adapted to discern which text is more difficult from two given texts (pairwise). As an exploratory study to prompt-learn a neural network for text readability in a text-to-text manner, we report useful tips for future work in seq2seq training and ranking-based approach to readability assessment. Specifically, we test nine input-output formats/prefixes and show that they can significantly influence the final model performance. Also, we argue that the combination of text-to-text training and pairwise ranking setup 1) enables leveraging multiple parallel text simplification data for teaching readability and 2) trains a neural model for the general concept of readability (therefore, better cross-domain generalization). At last, we report a 99.6% pairwise classification accuracy on Newsela and a 98.7% for OneStopEnglish, through a joint training approach.