Razvan Ciuca

• Razvan Ciuca 23 Oct 2017 0:35 UTC

  5 points

  in reply to: whpearson’s comment on: What Ev­i­dence Is AlphaGo Zero Re AGI Com­plex­ity?

  If Deep Learning people suddenly starting working hard on models with dynamic architectures who self-modify (i.e. a network outputs its own weight and architecture update for the next time-step) and they *don’t* see large improvements in task performance, I would take that as evidence against AGI going FOOM.

• Razvan Ciuca 22 Oct 2017 16:08 UTC

  9 points

  in reply to: John_Maxwell’s comment on: What Ev­i­dence Is AlphaGo Zero Re AGI Com­plex­ity?

  As far as I can see you can use the same techniques to learn to play any perfect information zero-sum game