Teichmann: H\"ormander conditions, transport equations and supervised learning.
Teichmann: We consider certain approximation problems from machine learning as controllability questions for (stochastic) transport equations. H\"ormander conditions lead to new viewpoints on universal approximation theorems. http://cvgmt.sns.it/seminar/684/
When
Fri Apr 5, 2019 9am – 10am Coordinated Universal Time
