ML Project for UWB Hackathon The goals of this project was to see how a machine will response to a member in the community that is infected
The Desired user experience is just watch and hopefully understand how the machine is avoiding infection
The Implmenetation was with Unity, Pytorch, Python, C# Unity was used to build the enviroment Pytorch was the engine for the agents, using Python C# was the scripts implemented into the enviroment to distigish a goal, infected person and the effects of running into each
The Learning algorithm that was used was only reinforcement Reinforcement is rewarding for the correct action and a punishment for the wrong
Lots of issues - physics engine, kool-aid man(people going through walls), Time constraint, Environment being too large for mass learning
What is next I would like to get the store, work, and park environment working for a better simulation, added an apartment setting, add options for the user to add objects or people. goals for different types of people
The User can't really interact with the current version. Future Goals
https://github.com/scrum-futch/VIrus-Simulator
For setting up and training your own models https://github.com/llSourcell/Unity_ML_Agents/blob/master/docs/installation.md
To Open the Project, Go to Assets -> Scenes -> Demo scene file Must have unity installed