29 - Neural machine translation via binary code prediction, with Graham Neubig

NLP Highlights

ACL 2017 paper, by Yusuke Oda and others (including Graham Neubig) at Nara Institute of Science and Technology (Graham is now at Carnegie Mellon University). Graham comes on to talk to us about neural machine translation generally, and about this ACL paper in particular. We spend the first half of the episode talking about major milestones in neural machine translation and why it is so much more effective than previous methods (spoiler: stronger language models help a lot). We then talk about the specifics of binary code prediction, how it's related to a hierarchical or class-factored softmax, and how to make it robust to off-by-one-bit errors. Paper link: https://www.semanticscholar.org/paper/Neural-Machine-Translation-via-Binary-Code-Predict-Oda-Arthur/bbedfd0380eb2e62f1c3b61aaf484d5867e6358d An example of the Language log posts that we discussed: http://languagelog.ldc.upenn.edu/nll/?p=33613 (there are many more).