Skip to Main content Skip to Navigation
Journal articles

Deep learning et authentification des textes

Étienne Brunet 1 Laurent Vanni 1, 2 
1 BCL, équipe Logométrie : corpus, traitements, modèles
BCL - Bases, Corpus, Langage (UMR 7320 - UCA / CNRS)
2 MAASAI - Modèles et algorithmes pour l’intelligence artificielle
CRISAM - Inria Sophia Antipolis - Méditerranée , UNS - Université Nice Sophia Antipolis (1965 - 2019), JAD - Laboratoire Jean Alexandre Dieudonné, Laboratoire I3S - SPARKS - Scalable and Pervasive softwARe and Knowledge Systems
Abstract : Using Deep Learning to attribute authorship of French literary texts While problems of attributing authorship or dating a text can be tackled using the usual methods of literary historians, it is equally possible to turn to statistical and computing tools. A range of intertextual measures have been proposed to describe variation within and across authors. To date no single method can claim an uncontested superiority comparable to the use of DNA in paternity suits or criminal investigations. The present study asks whether artificial intelligence may be able to play this role, and seeks the answer in research involving two corpora. The first concerns 20th century French literature: a deep learning algorithm is used on 50 texts by 25 authors (e.g., Roman Gary, Émile Ajar) with the goal of matching the two texts by the same author. Deep learning is perfectly accurate. The second corpus is drawn from French classical drama and here the algorithm also categorically distinguishes and matches plays by Racine, Corneille, and Molière. The only errors concern two plays (the French texts of Molière's Don Garcia of Navarre and Racine's The Litigants) where the comic genre takes precedence over authorial voice. This paper investigates the mechanisms of deep learning (with a more detailed treatment planned for a subsequent publication).
Complete list of metadata

Cited literature [6 references]  Display  Hide  Download
Contributor : etienne brunet Connect in order to contact the contributor
Submitted on : Saturday, May 2, 2020 - 10:44:10 PM
Last modification on : Sunday, June 26, 2022 - 2:49:30 AM


Files produced by the author(s)


  • HAL Id : hal-02561039, version 1



Étienne Brunet, Laurent Vanni. Deep learning et authentification des textes. Texto ! Textes et Cultures, Institut Ferdinand de Saussure, 2019, Texto! Textes et cultures, Volume XXIV, (n°1), pp.1-34. ⟨hal-02561039⟩



Record views


Files downloads