A functional architecture for autonomous driving (2015) [pdf]

kth.se

34 pointsamelius3 comments

Sometimes, when you define the internal structure of a system in a way that is more conceptually accessible for humans, it actually performs worse.

A simple model based RL system, like this (https://webdocs.cs.ualberta.ca/~sutton/book/ebook/figtmp63.p...), trained with enough data would implicitly recover all the necessary sub-systems.

Will it perform worse in this case, with this system, though? What is "enough data"? And how do you expect to get any sense of why the system made a particular decision if all you have is a distribution of probabilities from such a simplistic policy?

RL can in principle recover necessary subsystems, but you can't tell if that has actually happened. A human designed system may not be as comprehensive asymptotically as a learned system, but the designer can ensure that critical portions of the state space are covered.

Reply on news.ycombinator.com