RT @ml_hoedt: TIL that batch normalization actually has a predecessor that was published in tricks of the trade (1998!). https://t.co/kuU3…
493 followers
191 followers
TIL that batch normalization actually has a predecessor that was published in tricks of the trade (1998!). https://t.co/kuU3XkIwZ0 https://t.co/VxbDKhieT4
253 followers
Early Stopping - But When? | SpringerLink https://t.co/QOw5kEu9SM
1,864 followers
@markvanderwilk @nfusi @andrewgwils @hardmaru @planarrowspace Well, you can, it depends on what kind of metric you use: see paper by @ylecun team https://t.co/OzeM0WPRfE but I agree that GPs are not the same as kNN or linear regression. They are linear *sm