Loading...
Uploaded by szityu01 on May 28, 2008
An agent is trained by a reinforcement learning method of Szita & Lörincz to learn a value function for Tetris. See details in article at http://eotvos.elte.hu/~szityu/papers/SzitaLorincz05Learning.pdf
Science & Technology
Standard YouTube License
How did your represent the game to the learner?
taostoner1 3 months ago
@DVAFP That sounds like a great plan..... Just pit computers against each other with a set of rules and see who wins... Then throw in some people and see how they change.... :D
lzrd0113 9 months ago
@DVAFP - Probably, but since chess is a lot more complex than Tetris, it would take loads of time and computer resources.
cyborgtroy 1 year ago
what about chess? is it possible to teach a computer chess through reinforcement learning?
DVAFP 3 years ago
Load more suggestions
How did your represent the game to the learner?
taostoner1 3 months ago
@DVAFP That sounds like a great plan..... Just pit computers against each other with a set of rules and see who wins... Then throw in some people and see how they change.... :D
lzrd0113 9 months ago
@DVAFP - Probably, but since chess is a lot more complex than Tetris, it would take loads of time and computer resources.
cyborgtroy 1 year ago
what about chess? is it possible to teach a computer chess through reinforcement learning?
DVAFP 3 years ago