Improving sentence compression by learning to predict gaze

TL;DR

This paper introduces a multi-task learning approach using eye-tracking data to enhance sentence compression models, achieving performance comparable or superior to existing methods.

Contribution

It presents a novel multi-task learning algorithm leveraging eye-tracking corpora for improved sentence compression.

Findings

Performance is competitive with state-of-the-art methods.

Eye-tracking data improves sentence compression quality.

Multi-task learning enhances model robustness.

Abstract

We show how eye-tracking corpora can be used to improve sentence compression models, presenting a novel multi-task learning algorithm based on multi-layer LSTMs. We obtain performance competitive with or better than state-of-the-art approaches.