[Tutorial]: GameInterpreter

[edwardchalstrey1]

Describe the tutorial you want to see.

Add a tutorial demonstrating the "GameInterpreter" natural language game description to pygambit code research.

We likely want to demonstrate V2 of the framework, but we should also decide whether for subsequent versions the tutorial would be updated/replaced or a new one added, showing the progression of the methodology.

The notebook will act as a showcase of the research, and should demonstrate the GameInterpreter methodology interactively, linking to the paper for more details.

Who is the tutorial for?

The intended audience for this will be the same as the paper, so it would be ok to assume knowledge of game theory concepts. It also need not devote time to explaining the pygambit API, even if this is done in the paper, since other tutorials/docs already cover this.

Add links to papers and software.

• v2 code in private repo
• (v1 code for reference)

[rahulsavani]

GameInterpreter v1 code: https://github.com/zczlsde/GameInterpreter @rahulsavani is the v2 code also publicly available?

v2 is currently only in a private repo, but I have asked Shilong if he can add you to that.

[edwardchalstrey1]

Note: may be worth trying out deepmind's strategicwm approach in tandem since the code is open source. See the cultivate.ipynb notebook which takes the approach of growing a game tree node by node (they cite GameInterpreter v1 paper)