2
(Choose 1 answer)
A. True
B. False
When learning word embeddings, we create an artificial task of estimating P(target | context). It is okay if we do poorly on this artificial prediction task; the more important by-product of this task is that we learn a useful set of word embeddings.