Generative Adversarial Network (GAN) by Hand ✍️
Calculating AI by Hand: 15 of 28
Library › Calculating AI by Hand ✍️
Generative Adversarial Network (GAN) by Hand ✍️
Reinforcement Learning with Human Feedback (RLHF) by Hand ✍️
A Generative Adversarial Network (GAN) pits two networks against each other: a Generator that turns noise into fake data, and a Discriminator that learns to tell fake from real. Invented by Ian Goodfellow in 2014, GANs launched the field of generative AI for visual data.
Goal: Generate realistic 4-D data from 2-D noise.
Setup
Step 1 of 9: Given
4 noise vectors in 2D (N)
4 real data vectors in 4D (X)
Generator
Step 2 of 9: First Layer
Multiply the noise vectors with weights and biases to obtain new feature vectors
Step 3 of 9: ReLU
Apply the ReLU activation function, which has the effect of suppressing negative values. In this exercise, -1 and -2 are crossed out and set to 0.