This repo contains all the libraries and code necessary to replicate the experiments conducted in the paper : Semantic Utility. The folder that contains the code is SUA (Semantic Utility Agent)

The other folder is deprecated.

Installlation guide: The code requires the follow dependencies to be installed.

Python basic Libraries: pip install pickle pip install numpy pip install matplotlib

DeepLearning:

PyTorch:https://pytorch.org/

pip install torch

Environments:

GYM: https://github.com/openai/gym

pip install gym

ALE: https://github.com/mgbellemare/Arcade-Learning-Environment

pip install ale-py

For Atari roms make sure to download them (http://www.atarimania.com/rom_collection_archive_atari_2600_roms.html) and run:

ale-import-roms /path/to/roms/

Symbolic AI:

CNFgen:https://massimolauria.net/cnfgen/

pip3 install [--user] cnfgen

WMC/SDD library:https://pysdd.readthedocs.io/en/latest/

pip install pysdd

Sat solvers:https://github.com/pysathq/pysat

pip install python-sat[pblib,aiger]

Others:

Graph Analysis :https://networkx.org/documentation/stable/install.html

pip install networkx[default]