We study the problem of neural poem generation, focussing on Dante Alighieri, widely known for his Divine Comedy. This challenging task, where the machine has to capture the linguistic features that characterize the poet from the few available author’s works, it is addressed by using a syllable-based neural language model trained with a multi-stage procedure that exploits non-poetic works of the same author and also other publicly available huge modern Italian corpora.
Andrea Zugarini
PhD Student in Articial Intelligence
My research interests include language generation, information extraction, online learning, transfer learning.