Published November 2, 2019 | Version v1
Conference paper Open

Multi-Task Modeling of Phonographic Languages: Translating Middle Egyptian Hieroglyphs

  • 1. Computational Linguistics, Heidelberg University, Germany
  • 2. Computational Linguistics & IWR, Heidelberg University, Germany

Description

Machine translation of ancient languages faces a low-resource problem, caused by the limited amount of available textual source data and their translations. We present a multi-task modeling approach to translating Middle Egyptian that is inspired by recent successful approaches to multi-task learning in end-to-end speech translation. We leverage the phonographic aspect of the hieroglyphic writing system, and show that similar to multi-task learning of speech recognition and translation, joint learning and sharing of structural information between hieroglyph transcriptions, translations, and POS tagging can improve direct translation of hieroglyphs by several BLEU points, using a minimal amount of manual transcriptions.

Files

IWSLT2019_paper_2.pdf

Files (142.4 kB)

Name Size Download all
md5:00f27e13d51c7a31de7ee82ddfd5eed9
142.4 kB Preview Download