Nice glossary!, I miss some terms such as regularization, sequence embedding, optimization methods (ADAM, SADAM ...), transfer learning, one-shot learning, weight initialization, test time augmentation... but in general it covers few a lot of terms. I would collaborate anytime soon!
