Sign languages represent an interesting niche for statistical machine translation that is typically hampered by the scarce-ness of suitable data, and most papers in this area apply only a few, well-known techniques and do not adapt them to small-sized corpora. In this paper, we will propose new methods for common approaches like scaling factor opti-mization and alignment merging strategies which helped im-prove our baseline. We also conduct experiments with differ-ent decoders and employ state-of-the-art techniques like soft syntactic labels as well as trigger-based and discriminative word lexica and system combination. All methods are evalu-ated on one of the largest sign language corpora available. 1.
Philippe Dreuw, Johannes Förster, Yannick Gweth, Dan Joseph Stein, Hermann Ney, G. Martinez, Jaume Vergés Llahí, Onno Crasborn, Ellen Ormel, Wei Du, Thomas Hoyoux, Justus Piater, J. Miguel Moya, Mark Wheatley
Discussion(0)
No comments yet. Be the first to comment.