Séminaire

René Venegas - ICAR seminar- «From corpus linguistics to PEUMO: empirical evidence of a didactic intervention for thesis writing in engineering».

On The November 14, 2025

ENS de Lyon
15, Parvis René Descartes, Lyon 7
Zoom link available upon request from the organisers : denis.vigier[at] ens-lyon[dot]fr jean-philippe.mague[at]ens-lyon[dot]fr
3-5pm

® Glenn Carstens Peters (Unsplash)
® Glenn Carstens Peters (Unsplash)

René Venegas, researcher in residence at the Collegium, will present his research: ‘From corpus linguistics to PEUMO: empirical evidence of a didactic intervention for thesis writing in engineering ’ at the seminar of the Syntax, Meaning and Textualities team at the ICAR laboratory.

Academic writing is a cognitively demanding task, and there are few Spanish tools based on solid pedagogical frameworks to support it. This study validates PEUMO (Plataforma de Escritura Universitaria con Mediación Online), a corpus-linguistics- and genre-pedagogy-based tool designed to support engineering dissertation writing.
A six-session intervention comparing traditional teaching with the LC+PG approach implemented via PEUMO was conducted with 66 final-year computer engineering students in Chile. The tool combines automatic language processing techniques with BETO, a deep learning model that is applied to rhetorical analysis.
The results revealed significant differences in the quality of introductions and overall assessments, with the experimental group achieving a rating that was 7.3% higher. Almost half of the recorded revision episodes demonstrated the direct positive impact of the tool's suggestions.
These results confirm the effectiveness of PEUMO as a device for providing formative feedback and demonstrate its potential to improve academic writing in specific disciplines through the use of generative artificial intelligence.

René Venegas is a professor and researcher at Pontificia Universidad Católica de Valparaíso, specializing in academic discourse analysis and teaching, using computational and corpus linguistics.

ICAR laboratory website