In recent years automatic machine translation has dramatically improved in performance, with the availability of neural networks and huge amounts of data. Why does it work, and more importantly, when does it break? What kinds of language are machines terrible at translating, and is there anything we can do about it? What might change in the future?
Come along to learn which aspects of a Universal Translator seem relatively plausible, which seem next to impossible at the moment, and some machine learning explanations for why the Enterprise's computer might not convey the gory details when a Klingon is insulting you.