Concept information
Preferred term
negative sampling
Definition
- A technique used in word embedding models which involves randomly selecting negative examples to train the model to distinguish between positive and negative word pairs.
Broader concept
Example
- Training was done by applying negative sampling with 15 negative samples per target-context pair for 10 iterations over the entire corpus using stochastic gradient descent. (Komninos & Manandhar, 2016)
In other languages
URI
http://data.loterre.fr/ark:/67375/8LP-FJJS0XPD-0
{{label}}
{{#each values }} {{! loop through ConceptPropertyValue objects }}
{{#if prefLabel }}
{{/if}}
{{/each}}
{{#if notation }}{{ notation }} {{/if}}{{ prefLabel }}
{{#ifDifferentLabelLang lang }} ({{ lang }}){{/ifDifferentLabelLang}}
{{#if vocabName }}
{{ vocabName }}
{{/if}}