LESSWRONG
LW

All of danield's Comments + Replies

AlphaStar: Impressive for RL progress, not for AGI progress
danield70

New paper relevant to this discussion: https://arxiv.org/abs/1911.08265

Constructing agents with planning capabilities has long been one of the main challenges in the pursuit of artificial intelligence. Tree-based planning methods have enjoyed huge success in challenging domains, such as chess and Go, where a perfect simulator is available. However, in real-world problems the dynamics governing the environment are often complex and unknown. In this work we present the MuZero algorithm which, by combining a tree-based search with a learned model, achieves su
... (read more)
Reply
[AN #69] Stuart Russell's new book on why we need to replace the standard model of AI
danield90

Thanks for this summary / commentary, Rohin -- I found it helpful!

Reply
2Rohin Shah
Glad it was useful!
Machine Learning Projects on IDA
danield70

I'm excited about this. If you get any substantive feedback from people who take on these projects or decide not to, I'd be very interested to see a follow-up post.

Reply
2dxu
Seconded.
An Equilibrium of No Free Energy
danield50

I think this article / concept is incredibly useful, and singlehandedly justifies the existence of LW2. Thank you!

I want to go reread you and your research and see how the free energy concept could apply there -- if anyone else does, I'd love to hear thoughts.

Reply
5Ben Pace
I agree that the article does some incredibly useful conceptual work. I will say that I think Eliezer had written it and was planning to publish independantly of our LW 2.0 plans, and that I think the test of LW 2.0's success will be (in this case) the discussion following it, and more generally the communication between writers and thinkers from the whole community (things like this valuable critique of Eliezer's fire alarm post). Huh, I like the suggestion for applying this to Hamming's talk, I might do that myself and do a brief write-up. Thanks!