Idle observation: my (almost) 6-month old was able to figure out how to grasp a tiny ball cookie (1cm diameter) and put it in his mouth in ~5 attempts upon seeing the cookie for the first time. Extremely high problem complexity, extremely very low training sample size.
-
Show this thread
-
This is obvs not learning from scratch since he has been practicing manipulating objects for several hundreds of hours. however it *is* his first time manipulating something this small, which requires entirely different strategies. His previous objects had been more like ~5-15cm
1 reply 3 retweets 78 likesShow this thread -
Fascinating bit is that he tried a variety of approaches before converging towards a solution: try with one hand, try with the other hand, try to pick the object from the floor with his mouth directly (a completely different strategy!)
2 replies 3 retweets 76 likesShow this thread -
Replying to @fchollet
Curiosity driven learning with exploration <-> exploitation at play here. Not necessarily thru a reward signal. Wondering if this type of behavior is naturally emergent or interwoven in us since birth?
1 reply 0 retweets 0 likes
A baby's life only features extremely sparse negative/positive reinforcement signals (pain / hunger / food) and thus development is almost entirely intrinsically driven.
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.