« Pondération attentionnelle » : différence entre les versions
Aucun résumé des modifications |
Aucun résumé des modifications |
||
Ligne 4 : | Ligne 4 : | ||
Voir aussi '''[[Mécanisme d'attention]]''' | Voir aussi '''[[Mécanisme d'attention]]''' | ||
== Compléments == | |||
Nous proposons d'adopter le terme ''pondération attentionnelle'' qui est déjà en utilisation dans le domaine de la psychologie pour désigner un concept similaire. | |||
== Français == | == Français == | ||
''' | ''' pondération attentionnelle ''' | ||
''' pondération de l'attention ''' | |||
== Anglais == | == Anglais == | ||
Ligne 12 : | Ligne 15 : | ||
'''Attention Weights''' | '''Attention Weights''' | ||
Attention-weighting is a technique by which the model learns which part of the incoming sequence needs to be focused on. Think of it as the ‘Eye of Sauron’ scanning everything at all times and throwing light on the parts that are relevant. | <!-- Attention-weighting is a technique by which the model learns which part of the incoming sequence needs to be focused on. Think of it as the ‘Eye of Sauron’ scanning everything at all times and throwing light on the parts that are relevant. | ||
In natural language processing, this usually means assigning different levels of importance to different words in a sentence. It assigns importance to each word by calculating "soft" weights for the word's numerical representation, known as its embedding, within a specific section of the sentence called the context window to determine its importance. The calculation of these weights can occur simultaneously in models called transformers, or one by one in models known as recurrent neural networks. Unlike "hard" weights, which are predetermined and fixed during training, "soft" weights can | In natural language processing, this usually means assigning different levels of importance to different words in a sentence. It assigns importance to each word by calculating "soft" weights for the word's numerical representation, known as its embedding, within a specific section of the sentence called the context window to determine its importance. The calculation of these weights can occur simultaneously in models called transformers, or one by one in models known as recurrent neural networks. Unlike "hard" weights, which are predetermined and fixed during training, "soft" weights can --> | ||
== Source == | == Source == | ||
[https://doi.org/10.1146/annurev.psych.49.1.585 Source : Goldstone, R. L. (1998). ''Perceptual Learning. Annual Review of Psychology''] | |||
[https://arxiv.org/pdf/2211.07714 Source : arxiv] | [https://arxiv.org/pdf/2211.07714 Source : arxiv] | ||
[https://en.wikipedia.org/wiki/Attention_(machine_learning) Source: wiki] | [https://en.wikipedia.org/wiki/Attention_(machine_learning) Source: wiki] |
Version du 10 septembre 2024 à 13:52
Définition
Technique qui permet au modèle d'apprendre si une certaine partie de la séquence d'entrée nécessite une attention particulière.
Voir aussi Mécanisme d'attention
Compléments
Nous proposons d'adopter le terme pondération attentionnelle qui est déjà en utilisation dans le domaine de la psychologie pour désigner un concept similaire.
Français
pondération attentionnelle pondération de l'attention
Anglais
Attention-Weighting Attention Weights
Source
Source : Goldstone, R. L. (1998). Perceptual Learning. Annual Review of Psychology Source : arxiv Source: wiki
Contributeurs: Arianne , Patrick Drouin, wiki