Mesure de confiance de relation sémantique dans le cadre d'un modèle de langage sémantique
Catherine Kobus, Géraldine Damnati, Lionel Delphin-Poulat.

This article proposes a new confidence measure estimated for concept hypotheses given by a semantic language model. This confidence measure is based upon the ontology and the semantic relations linking concepts of a dialog application. It aims at measuring how high a concept hypothesis is related to the other hypotheses of an utterance. The semantic relation confidence measure is evaluated alone and in combination with a classical acoustic confidence measure. It is shown that the two confidence measures are complementary and yield good performance in terms of cross entropy relative reduction.