Machine Learning
At the University of Baltimore I took an opportunity during Practical Game Programming to go beyond the intended scope of the class for the final project and experiment with machine learning. I worked with the ML-Agents tool set available for Unity to develop some self-directing AI. The video below shows some of my experiments with machine learning as was presented at the end of the semester. I managed to create AI that learns both in the immediate environment and reacts and AI that learns, remembers (more or less), and is usable in later projects. I also began to experiment with more complex machine learning for a strategy defense game I was working on at the time. The intent was to have the player place defenses and simple entities that a machine-learning driven agent would try to out-maneuver to complete objects and sneak through a level. I managed to create an AI that could slowly learn areas to avoid to reach a destination, the first step to teaching it about line-of-sight and how to get around other AI entities that could detect them.