Fecha: 6 de Octubre de 2014 a las 12:00

Ponente: Ekaterina Ovchinnikova (Information Sciences Institute, USC, Los Angeles, USA)

Lugar de celebración: Sala 1.03, ETSI Informática, UNED (mapa)

Resumen:

In this talk, I will present a joint inference-based approach to natural
language understanding and scene recognition. The talk will cover such
general problems of inference-based frameworks as hypothesis selection
and discuss the solutions. In the last part of the talk, I will present
applications of a recent large-scale inference-based system to such
knowledge-intensive tasks as recognizing textual entailment, coreference
resolution, metaphor interpretation, and narrativization of videos.