
Egg-Bot Triumphs!
Check out how the Egg-Bot performed overnight and see the impact of splitting tasks into separate neural networks.

Hello Egg-Bot!
Building and training the Egg Bot, a new ML-Agent with separate neural networks for movement and turret control, in a zombie apocalypse scenario.

Obstacles Pose A Challenge
Evaluating the results of training the agents with obstacles as a new challenge.

Evaluating The Results
After our first night of training, it’s time to evaluate the results of our goal-seeking agent.

Let The Training Begin!
Training our agents on our procedurally generated platform using Proximal Policy Optimization (PPO).

Our Agents
A look into how we build our ML-Agents to learn in our procedurally generated environments.

A Procedurally Generated Platform
A Procedurally Generated Platform using a Hexagonal Grid Designed to Introduce Progressive Challenges for our Agents as They Progress.

“Whatever it is, the way you tell your story online can make all the difference.”