Date
Share

Dominique Brunato

Institute for Computational Linguistics “A. Zampolli”

Dominique Brunato is a researcher at the Institute for Computational Linguistics “A. Zampolli” of the National Council of Research in Pisa, where she is a member of the ItaliaNLP Lab research group. She has a background in theoretical and applied linguistics and a Ph.D. in Cognitive Science from the University of Siena. Her primary research lies at the intersection between theoretical and computational linguistics, with a specific interest in linguistic complexity, modeling language learning processes, and interpreting (neural) language models using linguistically- informed approaches.

Modeling Text Quality through a Computational Linguistics Perspective

What makes a text of good quality? How does the linguistic style of a text affect the reader’s perception of a text as well-written, engaging or easy to comprehend? Which measures are more reliable to automatically model such a multifaceted property of texts? In my talk I will provide an overview of factors involved in the assessment of text quality from a computational linguistics perspective, with a primary focus on linguistic complexity and coherence. By relying on the linguistic profiling framework (Biber, 1993; Val Halteren, 2004 among others), I will focus in particular on form-related properties and present some recent works which have been carried out in different domains and in a multilingual perspective using Profiling-UD (Brunato et al. 2020), a text analysis tool inspired to the principles of linguistic profiling that allows for the extraction of a large number of linguistic phenomena from text, included including those directly relevant to assessing text quality from the reader’s perspective.