Skip to Main content Skip to Navigation
Conference papers

Le deep learning comme défi pour identifier le style d'un écrivain : l'exemple de Jean Giono

Véronique Magri 1
1 BCL, équipe Logométrie : corpus, traitements, modèles
BCL - Bases, Corpus, Langage (UMR 7320 - UNS / CNRS)
Abstract : What if artificial intelligence could identify a writer's style ? What if, automatically, the machine could identify the characteristics of a piece of writing, in other words the formal elements recognizable from one work to another, as well as the differences between a corpus of study and a reference corpus ? If finally a writing could be deciphered by an algorithm ? This is precisely the challenge of deep learning applied to literature. This is exactly the experimentation that is being attempted on Giono, from an unpublished digital text base, a very large corpus that brings together Giono's novels. The necessary differential measurement guides the constitution of the corpus ; two bases were thus constituted by É. Brunet : one brings together Giono's works, processed by the Hyperbase software. The other is a vast reference corpus whose generic and temporal homogeneity is guaranteed since they are 50 novels from the twentieth to the twenty-first centuries. The corpus was constituted by É. Brunet and includes two texts written by the same author, i.e. 50 texts for 25 authors. From prediction to deconvolution, an interpretative path is built, tending towards the horizon of the discovery of an author's style.
Complete list of metadatas

Cited literature [10 references]  Display  Hide  Download
Contributor : Véronique Magri <>
Submitted on : Friday, September 11, 2020 - 11:57:42 AM
Last modification on : Wednesday, September 16, 2020 - 3:32:41 AM


Files produced by the author(s)


  • HAL Id : hal-02936437, version 1



Véronique Magri. Le deep learning comme défi pour identifier le style d'un écrivain : l'exemple de Jean Giono. JADT, Jun 2020, Toulouse, France. ⟨hal-02936437⟩



Record views


Files downloads