Deep learning languages: a key fundamental shift from probabilities to weights? - Archive ouverte HAL Access content directly
Preprints, Working Papers, ... Year :

Deep learning languages: a key fundamental shift from probabilities to weights?

(1)
1

Abstract

Recent successes in language modeling, notably with deep learning methods, coincide with a shift from probabilistic to weighted representations. We raise here the question of the importance of this evolution, in the light of the practical limitations of a classical and simple probabilistic modeling approach for the classification of protein sequences and in relation to the need for principled methods to learn non-probabilistic models.
Fichier principal
Vignette du fichier
probxit.pdf (185.95 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-02235207 , version 1 (01-08-2019)

Identifiers

Cite

François Coste. Deep learning languages: a key fundamental shift from probabilities to weights?. 2019. ⟨hal-02235207⟩
63 View
137 Download

Altmetric

Share

Gmail Facebook Twitter LinkedIn More