no code implementations • 23 Oct 2019 • Jun Liu, Jiedan Zhu, Vishal Kathuria, Fuchun Peng
A second layer is a private cache that caches the graph that represents the personalized language model, which is only shared by the utterances from a particular user.