Google published a pretty detailed research article about their approach a year ago. [1] A competent deep-learning team with access to a compute farm of ~1000 g

Google published a pretty detailed research article about their approach a year ago. [1] A competent deep-learning team with access to a compute farm of ~1000 gpu's could be be following the google recipe. That is a nontrivial expense though, e.g. not some college kid in their dorm room. [1] http://www.nature.com/nature/journal/v529/n7587/full/nature16961.html http://www.nature.com/nature/journal/v529/n7587/full/nature1...