Robust Automatic Metaphors Development with Deep Learning

Robust Automatic Metaphors Development with Deep Learning

Institute(s): TU Darmstadt

Associated since: December 2015project participants

Prof. Dr. Iryna Gurevych, TU Darmstadt
Erik-Lân Do Dinh, TU Darmstadt

CEDIFOR partner:

Prof. Dr. Petra Gehring, Institute for Philosophy, TU Darmstadt
Prof. Dr. Andrea Rapp, Institute for Linguistics and Literature, TU Darmstadt

Brief description

Continuation of the pilot project “Computer-aided indexing of metaphors in theory-language educational text corpora”. A more robust solution for automated metaphor detection is currently being worked on, which, in contrast to conventional methods, requires a neural network and word embeddings without further background corpora.
Resources

Associated Software Products

CSniper
DKPro Core

Publications:

Do Dinh, Erik-Lân and Gurevych, Iryna: Token-Level Metaphor Detection using Neural Networks. In Proceedings of The Fourth Workshop on Metaphor in NLP held in conjunction with NAACL 2016, p. 28-33, 2016.