Single-Layer Perceptron in Rust

What You'll Build

We'll build a tiny but complete learning machine: a single-layer perceptron packaged as a reusable Rust library crate. Starting from an empty project, we'll define the data structure, initialize random weights, implement the forward pass and training loop, then verify everything with integration tests for AND, OR, and XOR logic gates.

Along the way we'll see how the perceptron learning rule adjusts weights, why convergence is guaranteed for linearly separable problems, and what happens when a problem (XOR) isn't linearly separable.

Learning Objectives

Set up a Rust library crate from scratch
Define a generic perceptron struct with const generics
Initialize weights and bias randomly with rand
Implement the forward pass with weighted sum and step activation
Implement the perceptron learning rule with early stopping
Write integration tests for AND, OR, and XOR logic gates
Understand the linear separability limitation of single-layer perceptrons

Prerequisites

Basic Rust syntax (structs, impl blocks, generics)
Rust toolchain and cargo installed
Comfort with arrays, slices, and iterators
Basic understanding of binary classification

Course curriculum

1 Configurable Perceptron Weights

Initialize Perceptron Library
Add Perceptron Skeleton
Randomize Initial Weights

2 Perceptron Training Logic

Add Training Loop
Compute And Train Perceptron
Configure Learning Rate
Stop Training Early

3 Prediction Method

Add Prediction Method

4 Training Logic Gates

Train AND Perceptron Example

5 Training Logical OR

Train OR Perceptron

6 Perceptron Limits with XOR

Train And Test XOR

Technologies

Rust Perceptron Machine Learning Binary Classification Activation Function Learning Rate Weights and Bias Linear Separability Const Generics rand

Read the concepts behind this playground

FAQ

Is this a video course?

No. You implement the project yourself, step by step, right in your browser — with a full reference solution for every step when you get stuck.

Do I need to install anything?

No. The playground runs entirely in your browser — no toolchain or local setup required.

Can I try it before subscribing?

Yes. Preview steps are available without a subscription so you can see the format and quality first.

What do I get with full access?

Every step unlocked, the Workbench app, a snapshot for each step, jump-to-any-step navigation, and all new playground releases.

Does it work with an AI agent like Claude or Codex?

Yes. You can work through a playground solo or alongside an AI agent.

Is there reference code?

Yes. Every step has a complete reference solution you can compare against.