Implicit Regularization I

preview_player
Показать описание
Nati Srebro (Toyota Technological Institute at Chicago)
Deep Learning Boot Camp
Рекомендации по теме
Комментарии
Автор

At 43:45, I don't see why that's a quadratic. I think all the terms on the right side change linearly, which would mean there would never be a min. The only way I see for it to be quadratic is if that was the norm squared, not just the norm

grantsmith