Skip to content

Latest commit

 

History

History
19 lines (14 loc) · 687 Bytes

README.md

File metadata and controls

19 lines (14 loc) · 687 Bytes

We are going to create agents for decision-making in deep and complex sequential environments with huge state spaces,
like Pluribus which beats top professional poker players in No-Limit Holdem poker with 10^161 states. We can do this
with combining following three main components

CFR Algorithms

For making decisions in environments with sequential structures, like games, specifically imperfect information games

  • We are going ti implement different variations of CFR
    • Vanilla CFR
    • CFR+
    • Chance Sampling CFR
    • Linear CFR
    • MCCFR
    • Our improvements of CFR

Real time search and Depth Limited Solving in subgames

State space clustering and abstractions