This is a rough draft implementation that was made by Matt Jacobs with the below resources and their associated credits,
Credits:
This implementation was made using Unity and it's machine learning add-on known as ML-Agents. Their website can be found at: https://unity.com/
The code pertaining to the movement/actions of the cubes , the hyper-parameter set up for the reinforcement learning, and the code for cubes making observations for the neural network is from the penguin tutorial found at immersivelimit.com . All code that was used from them has been commented . Here is the link to the penguin tutorial on their website : https://www.immersivelimit.com/tutorials/reinforcement-learning-penguins-part-1-unity-ml-agents
The tower game object was downloaded for free from the Unity game store. It was made by onpolyx and their website can be found at: http://www.onpolyx.com/
The emojis used for facial expressions are from Loren Sands-Ramshaw at Medium.com and can be found at:
The food objects were downloaded for free from the Unity game store. The package name is "Frood" and they were made by voxelguy and his website can be found at: https://www.voxelguy.fr/
The theory of humor that this implementation is based on is called the Benign Violation Theory and was created by McGraw. His book, The Humor Code: A Global Search for What Makes Things Funny, can be found for sale at: https://www.amazon.com/Humor-Code-Global-Search-Things/dp/1451665423/ref=sr_1_1?dchild=1&keywords=the+humor+code&qid=1591577515&sr=8-1