Creating curious agents that can explore their environment and actions without obvious rewards seems like it'll lead to dramatic progress.
It's an old idea, but a powerful one. @pyoudeyer has been working on this since the mid-2000s http://www.pyoudeyer.com/curiosity-and-information-seeking-in-cognitive-development/ …
-
-
Awesome, thanks for the link!
-
but we all know who did it in 1991 though
- Show replies
New conversation -
-
-
I worked on this idea as well in 2010-2012, inspired by
@pyoudeyer's papers -
Did you publish anything on it?
- Show replies
New conversation -
-
-
yeah if you look for info. maximization or uncertainty reduction more work comes up. This is from our lab: http://ieeexplore.ieee.org/document/7324371/ …pic.twitter.com/IflhudGwEC
-
if we consider reducing the uncertainty about and environment as "acting for the curiosity to learn more" we unify both approaches/syntaxes
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.