When we design hybrid systems which combine machine-learning algorithms with symbolic language resources elaborated by linguists, we divide tasks between linguists and algorithms, and we select models that facilitate this division of tasks. What are the challenges faced by linguists and by machine learning algorithms in this context? What are the assets that allow them to take on these challenges? Is the problem different for different types of resources? Which trends in linguistics equips linguists best for building resources?
from HAL : Dernières publications http://ift.tt/19RVwVA
from HAL : Dernières publications http://ift.tt/19RVwVA

0 commentaires:
Enregistrer un commentaire