Four: Brian Christian on artificial intelligence
Effective Altruism: Ten Global Problems – 80000 Hours
00:00
The Deep Cued Networks Part of Machine Learning
D c n uses neural networks as a function approximator for cue learning. Cu learning is a type of value learning, which is a way to solve reinforcement learning. So when we've now popped pop to the stack, we're back to where we started. Ok, totally clear? Ha, ha. We hope it's somewhat clear. All right. Djus thing of it as going like in one direction only. And i and i guess not upwards, or skipping a level or rare. There are some architectures that have cycles in them, like l s t m are what's called recurrent, but let's just pretend that doesn't exist.
Transcript
Play full episode
