Linear Algebra Is the Language Neural Networks Speak

What Actually Happens in model.forward(x)?

When you call model(x) in PyTorch:

output = activation(W @ x + b)

That @ symbol? Matrix multiplication. Every. Single. Layer.

Vectors Are Data

In AI, everything is a vector: words, pixels, preferences.

Matrices Are Transformations

A weight matrix transforms space — rotating, scaling, and reflecting data in high-dimensional space.