Learning and improving backgammon strategyGreg Galperin
Web page, from Center for Biological and Computational Learning. Proposes using rollouts to learn the evaluation function, for example, by training a neural network.
Greg Galperin <grg@ai.mit.edu> noticed that rollouts can be done in parallel: you can use more than one processor to play positions out with virtually no loss of efficiency. So he's been investigating the use of powerful parallel computers to play backgammon.
| Jay : game learning : backgammon : rollout |
With a fast enough computer, it's possible to roll a position out in real time, creating a program that could be much stronger than current programs.
Learning and improving backgammon strategy
Greg Galperin
Web page, from
Center for Biological and Computational Learning.
Proposes using rollouts to learn the evaluation function,
for example, by training a neural network.
Machine learning for prediction and control (2 pages)
Greg Galperin and Paul Viola
Postscript, from
MIT AI Lab Learning & Vision Group.
Also available in
PDF format.
This seems to be an updated version of the above; it describes
the project without giving details.
Parallel Monte-Carlo simulation of neural network controllers
Greg Galperin and Gerald Tesauro
Web page, from
MCS Division, Argonne National Laboratory.
Proposes using rollouts to play games in real time.
They're planning to create a program clearly stronger than any human.
(This one seems to have gone away. Anybody have an alternate source?)