MoEAtt: A Deep Mixture of Experts Model using Attention-based Routing Gate.

Gal Blecher, Shai Fine

Browse the full ICMLA paper archive.