Concept information
Terme préférentiel
label smoothing
Définition
- A technique used in neural networks for tasks like image classification and speech recognition to prevent model overfitting and improve its generalization ability by distributing probability mass to all labels, not just the gold label.
Concept générique
Exemple
- During the training process the value of label smoothing is set to 0.1 and dropout is set to 0.1. (Cheng, Yao, Xin, An, Li, Li & Zou, 2024)
Traductions
-
français
URI
http://data.loterre.fr/ark:/67375/8LP-S7V1DRMJ-9
{{label}}
{{#each values }} {{! loop through ConceptPropertyValue objects }}
{{#if prefLabel }}
{{/if}}
{{/each}}
{{#if notation }}{{ notation }} {{/if}}{{ prefLabel }}
{{#ifDifferentLabelLang lang }} ({{ lang }}){{/ifDifferentLabelLang}}
{{#if vocabName }}
{{ vocabName }}
{{/if}}