C++
OCT 2023


During my first quarter at UCLA, I took a 2-unit course on basic neural network systems. I initially found the math behind it to be overly complex, but after taking multivariable calculus I realized it was actually quite simple when broken down, so I decided to build my own neural network library.

The neural network takes in the pixel coordinates as inputs and three RGB values as the expected outputs. It trains in real time on the variously colored dots inputted by the user.

The training iterations are sped up with multithreading implemented with OpenMP.