Re:Visual Q-Learning Demo in JavaScript from Karpathy (2013)
by throwaway4aday on Saturday November 25, 2023 @03:07PM
(#38416248)
Sounds like the process of tuning the reward space is a type of labelling and ranking problem. If I'm not mistaken, those are two things that GPT-4 is pretty good at. You wouldn't even necessarily pre-label every possible action since GPT-4 could do it in real time.
[ Parent
| on HN ]
There are no comments on this comment yet.