Automatic design of semantic similarity controllers based on fuzzy logics Academic Article uri icon

abstract

  • As for the calculation of semantic textual similarity, the use of neural networks to vectorize words suffers from interpretability, that is, the capacity that a person has to understand the model perfectly. For this reason, semantic similarity controllers based on fuzzy logics have appeared, which can explain the model based on rules close to natural language easily understandable by a person.

publication date

  • January 1, 2019