Text Simplification Beyond Sentence-Level

Abstract

Text Simplification aims to modify a text to make it easier to read. Current approaches focus on sentence-level simplifications, mostly because of unavailability of appropriate data. A recently released corpus, created by professional editors, allows us now to study simplifications at the document-level, and develop methods able to perform rewriting transformations not restricted to sentence boundaries. During the poster session, I would discuss our findings related to text simplification with a document-level perspective, and how we intend to devise novel corpus-based simplification methods, so that they are able to perform cross-sentence operations. I will review our results on a manual analysis of professionally produced simplifications, which sheds light on the transformation operations that humans perform at the document-level. Also, I will describe our work-in-progress on automatically identifying these operations, and how this information can help in the implementation of text simplification systems, and in a more operation-based evaluation of simplification performance.

Date
Oct 31, 2017 17:00
Event
5th Annual Amazon Graduate Research Symposium
Location
Amazon
2021 7th Avenue, Seattle, WA 98121
Fernando Alva-Manchego
Fernando Alva-Manchego
Research Associate

My research interests include text simplification, readability assessment, evaluation of natural language generation, and writing assistance.