[24] Martin Arjovsky - Out of Distribution Generalization in Machine Learning
The Thesis Review
Wasterstein Gan - Was It Really a Novel Approach?
The theory was finished before the first experiments and in the these two papers I was very lucky that I I knew Sumit very well from my time in Facebook and I knew how the experiments were gonna turn out because he had worked a lot on guns to know how they behave. The thing that was that took a lot of time or I guess most of the time in was the experiment design part like saying okay what do we need to do like what are the rights experiments for these papersWhat are our claims and how to validate those claims with experiments? And this is how Sumit came up with like training network without batch norm and like with a different discriminator.
00:00
Transcript
Play full episode
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
