Passer au contenu principal

Vocabulary of natural language processing (POC)

Choisissez le vocabulaire dans lequel chercher

Concept information

Terme préférentiel

metric learning  

Définition

  • A technique that constructs task-specific distance metrics by learning the similarity or dissimilarity between samples. It is often used for object retrieval and clustering by training a deep neural network to learn a mapping function from the original samples into a new, more compact, embedding space. (Seidenschwarz et al., Learning Intra-Batch Connections for Deep Metric Learning)

Concept générique

Synonyme(s)

  • distance metric learning

Traductions

URI

http://data.loterre.fr/ark:/67375/8LP-KR0VQCZP-S

Télécharger ce concept :

RDF/XML TURTLE JSON-LD Dernière modification le 28/05/2024