Generalized Model Learning for Reinforcement Learning on a Humanoid Robot

[privatepile762]

B.F. Skinner would be proud! Finally, an extension of his science to other aspects of humanity, specifically the development of artificial intelligence. What an elegant curve in the behavior of the robot, too! Great work. Perhaps other studies on variable, fixed, and interval schedules of reinforcement will yield data trends consistent with living organisms?

[fragsman]

I might do a research about this for my last subject before I get my degree on computer science. It will be hard but at least i'm near the end :p

[fragsman]

Nice, he learns through rewards.

[roleplayman]

so the RL is to do random actions when the reward is minus and when it maximizes you improve by small steps [IE Learning Rate] ?

[amightyo]

RL is Nao!!!!