
Episode 15: Martín Arjovsky, INRIA, on benchmarks for robustness and geometric information theory
Generally Intelligent
00:00
We Need More Benchmarks on Robastness
We need more of this. We kneel o at scale, because these problems exist. I would just like to have bencmarkd things were, i contest aorisens on this. There are so many edge cases all the time that it is difficult to continue building existing systems and get something with really good uptianization out of them. If this was my fultam job, what i would do is i would go to startups that are working on agriculture,. Thatis the first lebthat i would take, because these are the people whose problems you want to solve. And the bench pruks come as a model for that. Because self driving cars is one of the
Transcript
Play full episode