This week I went more in-depth researching advanced poker algorithms and their implementation, and made decisions on what I should actually implement. The goal is to have a model that can be competitive vs casual or amateur poker players, not necessarily to be able to beat professionals. Thus, the basic idea is to take portions of the Libratus and Pluribus papers and either simplify or implement features in a less granular way, so that the model can both be simpler to implement and take less time to train. At the moment, the simplifications are likely to be a much less refined subgame solving protocol similar to the Pluribus one, as well as rounding bet amounts, and other simplifications. I am on schedule, and next week I will be looking to implement the base of the poker game on which I can begin to implement monte carlo methods.
https://www.science.org/doi/10.1126/science.aao1733
https://www.science.org/doi/10.1126/science.aay2400