All posts
Knowledge AI ML April 9, 2026

ReLU, Neural Nets, and Backpropagation

A compact sketch of ReLU activations, neural network structure, and backpropagation flow.

This sketch is a compact overview of the pieces I want visible when thinking about basic neural networks: layered structure, ReLU activations, the forward pass, loss calculation, and how backpropagation pushes gradients back through the model.

ReLU, Neural Nets, and Backpropagation

What It Covers

  • how a simple feedforward neural network is organized
  • where ReLU sits in the forward pass
  • how activations move layer by layer toward an output
  • how loss connects to gradient flow during backpropagation
  • a high-level mental model for parameter updates

It is the kind of note I want nearby before going deeper into implementation details, training behavior, or optimization tradeoffs.

Thanks for reading.

© 2026 Alan Wang