« Autoencodeurs peu denses » : différence entre les versions
(Page créée avec « ==en construction== == Définition == XXXXXXXXX == Français == ''' XXXXXXXXX ''' == Anglais == ''' Sparse Auto Encoder''' In sparse autoencoders, we can still use f... ») |
Aucun résumé des modifications |
||
| Ligne 12 : | Ligne 12 : | ||
In sparse autoencoders, we can still use fully connected neurons with numbers equal to the image dimensionality. But still, by adding a sparsity regularization, we will be able to stop the neural network from copying the input. | In sparse autoencoders, we can still use fully connected neurons with numbers equal to the image dimensionality. But still, by adding a sparsity regularization, we will be able to stop the neural network from copying the input. | ||
Mainly, there are two ways to add sparsity constraints to deep autoencoders. | Mainly, there are two ways to add sparsity constraints to deep autoencoders. | ||
= L1 regularization, which we will use in this article. | = L1 regularization, which we will use in this article. | ||
Version du 23 octobre 2023 à 09:13
en construction
Définition
XXXXXXXXX
Français
XXXXXXXXX
Anglais
Sparse Auto Encoder
In sparse autoencoders, we can still use fully connected neurons with numbers equal to the image dimensionality. But still, by adding a sparsity regularization, we will be able to stop the neural network from copying the input. Mainly, there are two ways to add sparsity constraints to deep autoencoders. = L1 regularization, which we will use in this article. * KL divergence, which we will address in the next article.
Contributeurs: Marie Alfaro, wiki
