Analyse sémantique stochastique

Anglais

Stochastic semantic analysis

Stochastic semantic analysis is an approach used in computer science as a semantic component of natural language understanding.

Stochastic models generally use the definition of segments of words as basic semantic units for the semantic models, and in some cases involve a two layered approach.[1]

Example applications have a wide range. In machine translation, it has been applied to the translation of spontaneous conversational speech among different languages.[2] In the area of spoken language understanding the fact that spoken sentences often do not follow the grammar of a language and involve self-corrections, repetitions, and other irregularities, the use of stochastic semantic has been suggested as a natural fit to achieve robustness to deal with noise due to the spontaneous nature of spoken language.[3]