(1/n) This week we have
@fredsala
on the Stanford MLSys Seminar, live on Thursday at 1:30 PM! Fred was a postdoc at
@StanfordAILab
, and is now a professor at
@WisconsinCS
and a research scientist at
@SnorkelAI
-- so he knows a thing or two about MLSys.
(2/n) Fred's talk will be about efficiently constructing datasets for diverse datatypes. He'll cover a new technique for fusing weak supervision with structured prediction. This allows labels that can be continuous, "manifold-valued," sequences, graphs, and more!
(3/n) As always, Fred's work is a healthy mix of theory and experiments. He'll discuss theoretical guarantees that draw connections to embeddings of metric spaces, and show experimental results in problems ranging from ranking to geodesic regression.