In this second assignment I trained a neural network to recognize different guitar body types. What was produced was an infinite matrix of "hybrids" between these four. Building upon our understanding of vector-based semantic space, I could "multiply" these forms again by other formal semantic spaces.

I demonstrated the capability to create variations through a grasshopper script. These variations would then be read by a computer vision python script, which would "see" objects held up to a webcamera, and cateogrized them based on training data.

In my eventual fabricated object, I extracted different hybrids generated through varying depths of multiplicative hybridization. Recognizing the size limits of a 3D printer, I took this as an opportunity to represent these differences by creating 4 splits in the guitar body where these forms could intersect with each other. They would all connect to a 3D-printed guitar body core, as well as to the neck made for my telecaster, which could also be used here.