This repository contains the Game AI framework based on the popular board game Splendor designed by Marc André and published by Space Cowboys. The engine allows the user to modify all the game design parameters to play a wide variety of Splendor-like games, but also to create new game content and develop new game-playing agents. Rinascimento intends to promote the development and benchmarking of several applications of Game AI, few examples being: game-playing, Procedural Content Generation, automated game design, agent hyperparameter tuning.
For more details you can read the material linked below.
If you are interested in running some experiments don't hesitate to contact me at my QMUL email address, I'm happy to help and support your research with Rinascimento!
Bear in mind this is being developed as an accademic project just for research purposes.
Wiki and docs coming soon(-ish)!
WARNING! Since April 2023 I've uploaded some datasets that require LFS, make sure to install it following the relative instructions.
This repository contains a set of experiments and scripts I developed during my IGGI PhD. Here's a list:
- Play Rinascimento games between AIs and/or user input with GUI;
- Hyperparameter tuning using NTBEA or grid search algorithms;
- Behavioural Search using the MAP-Elites search (paper out soon);
Rinascimento: Optimising Statistical Forward Planning Agents for Playing Splendor - paper - talk - slides
Game AI hyperparameter tuning in rinascimento - paper
Rinascimento: using event-value functions for playing Splendor - paper - talk - slides
Being an Italian word I understand people may have some issue pronouncing it.
Don't panic, it's ok if you butcher it!
But if you wanted to learn here's how you would pronounce it if you were an English speaker: reenah-she-ment-o
And if you haven't given up yet, here's a YouTube video
You can send my an email or follow me on Twiter.
If you're interested in my research visit my Google Scholar profile.