I am excited to announce ALEx, the (A)rcade (L)earning (E)nvironment for Eli(x)ir.
The ALE is a framework for writing AI agents for Atari 2600 games. It uses the Stella Emulator to emulate Atari ROMs. You can use the ALE to interact with the game environment.
You can see an example of an ALEx agent playing Tetris here:
ALEx exposes the ALE Interface with NIFs and lets you interact with the Screen, State, RAM, and do a bunch of other cool things. Admittedly, I’m not the best at C/C++, so I’m hoping some of you out there would want to help contribute!
I wrote ALEx for a Reinforcement Learning Library I’m currently developing. Stay tuned!