Skip to Main content Skip to Navigation
Conference papers

ADT et deep learning, regards croisés. Phrases-clefs, motifs et nouveaux observables

Abstract : This contribution confronts ADT and Machine learning. The extraction of key-statistical passages is first proposed according to several calculations implemented in the Hyperbase software. An evaluation of these calculations according to the filters applied (taking into account of the positive specificities only and substantives only, etc.) is given. The extraction of key passages obtained by deep learning-passages that have the best recognition rate at the time of a prediction-is then proposed. The hypothesis is that deep learning is of course sensitive to the linguistic units on which the computation of the key statistical sentences are based, but also sensitive to other phenomena than frequency and other complex linguistic observables that the ADT has more difficult to take into account-as would be linguistic pattern (Mellet et Longrée, 2009). If this hypothesis is confirmed, it would on the one hand better apprehend the black box of deep learning algorithms and on the other hand to offer the ADT community a new points of view. Résumé Cette contribution confronte ADT et Deep learning.
Complete list of metadata

Cited literature [11 references]  Display  Hide  Download
Contributor : Damon Mayaffre Connect in order to contact the contributor
Submitted on : Tuesday, June 26, 2018 - 10:50:33 AM
Last modification on : Saturday, June 25, 2022 - 11:30:59 PM
Long-term archiving on: : Wednesday, September 26, 2018 - 8:38:50 PM


Files produced by the author(s)


  • HAL Id : hal-01823560, version 1



Laurent Vanni, Damon Mayaffre, Dominique Longrée. ADT et deep learning, regards croisés. Phrases-clefs, motifs et nouveaux observables. JADT 2018, Jun 2018, Rome, Italie. ⟨hal-01823560⟩



Record views


Files downloads