Foundations of AI

An overview of Rosenblatt’s 1958 perceptron — how it works, where it fails, and its influence on modern neural networks.

Activation functions explained: their role in neural network learning, evolution from step functions to ReLU and beyond, and why non‑linearity enables deep learning.

MLPs and backpropagation: how deep networks learn by stacking non-linear layers—and where they struggle.