Episode 13: Jonathan Frankle, MIT, on the lottery ticket hypothesis and the science of deep learning
Generally Intelligent
00:00
Is There a Way to Start Small?
The author of the original 'lotto ticket' paper on c far ten. He says he had no resources and struggled to get access to them. The results fall apart if you do this on immagene, but they don't work on normal networks for c far 10. But working on emnist can give you a sense that maybe something might have some legs.
Transcript
Play full episode
