Difference between revisions of "Batch Learning for Poker"

From AIRWiki
Jump to: navigation, search
(Project short description)
Line 14: Line 14:
  
 
This project is aimed at determination an efficient policy to play Poker Limit Hold`em Heads up. Subdivided in three major areas:
 
This project is aimed at determination an efficient policy to play Poker Limit Hold`em Heads up. Subdivided in three major areas:
1- Analysis of rules of poker and player strategies.
+
  1- Analysis of rules of poker and player strategies.
 
2- Feature Extraction: Given the fact that number of possible states are quite large (10E14). A FS approach is necessary.
 
2- Feature Extraction: Given the fact that number of possible states are quite large (10E14). A FS approach is necessary.
 
3- Fitted Q-Iteration to determinate the policy using Extremely Randomized Trees.
 
3- Fitted Q-Iteration to determinate the policy using Extremely Randomized Trees.

Revision as of 13:22, 15 May 2010

MRT: Batch Learning for Poker
Coordinator: MarcelloRestelli (restelli@elet.polimi.it)
Tutor:
Collaborator:
Students: RafaelVilella (rafaelcarioca51@gmail.com)
Research Area: Machine Learning
Research Topic:
Start: 2010/03/01
End: In progresswarning.pngThe date "In progress" was not understood.
Status: Active
Type: Thesis


Project short description

This project is aimed at determination an efficient policy to play Poker Limit Hold`em Heads up. Subdivided in three major areas:

 1- Analysis of rules of poker and player strategies.

2- Feature Extraction: Given the fact that number of possible states are quite large (10E14). A FS approach is necessary. 3- Fitted Q-Iteration to determinate the policy using Extremely Randomized Trees.

Dates

Start date: 2010/03/01

End date: Still in progress

Project head(s)

M.Restelli - User:MarcelloRestelli

Students currently working on the project

Rafael Domingues Santos Vilella - User:RafaelVilella

Laboratory work and risk analysis

This project is related to software developing so there are no dangerous activities