|
@OpenAI | |||||
|
We've observed AIs discovering complex tool use while competing in a simple game of hide-and-seek. They develop a series of six distinct strategies and counterstrategies, ultimately using tools in the environment to break our simulated physics: openai.com/blog/emergent-…
|
||||||
|
||||||
|
OpenAI
@OpenAI
|
17. ruj |
|
Unexpected and surprising behaviors included box surfing, where seekers learn to bring a box to a locked ramp in order to jump on top of the box and then “surf” it to the hider’s shelter. pic.twitter.com/v0kGfCYZna
|
||
|
|
||
|
OpenAI
@OpenAI
|
17. ruj |
|
And seekers learn that if they run at a wall with a ramp at the right angle, they can launch themselves upward. pic.twitter.com/SJv9SzctEp
|
||
|
|
||
|
hardmaru
@hardmaru
|
18. ruj |
|
Congrats on this very inspiring work! I'd be interested to see this line of work scale up in a way proposed in your previous paper on Neural MMO paper: twitter.com/OpenAI/status/…
|
||
|
|
||
|
Joseph Suarez
@jsuarez5341
|
18. ruj |
|
I desire nothing more than the opportunity to do so. Have been hard at work on Neural MMO since the v1.0 release -- lots of progress since: pic.twitter.com/PXvkEEK2XP
|
||
|
|
||
|
H2O Delirious
@H2ODelirious
|
19. ruj |
|
Always Red Vs Blue LOL This is actually amazing to see them learn techniques to one up each other.
|
||
|
|
||
|
Jakob Foerster
@j_foerst
|
17. ruj |
|
This is great and looks really cool. Any ideas for how to compare this to complex sequences of moves emerging in other self-play settings such as Go or Dota?
|
||
|
|
||
|
Amardeep Singh Sidhu
@adssidhu86
|
17. ruj |
|
Hahaha Hiders taking away all the blocks and ramp is my favourite part🤣🤣🤣
|
||
|
|
||
|
Adam
@Dobra_Jest
|
18. ruj |
|
The best was block surfing !
|
||
|
|
||