Teaching Thread Part B: Multi-Class Perceptron Rule Derived on the Board

alfcnz · x · 2026-09-24

Part B of @alfcnz's teaching thread derives the multi-class perceptron: from s = wᵀx with w ← w + (y − ỹ)x, extending to competing directions s = Wx where the prediction is one-hot on the highest score, with update rule W ← W + (y − ỹ)xᵀ — the same formula with lists. Includes a worked example of one wrong guess.

Related event: Neural Network Tutorial Thread: From Perceptrons to Multilayer Representations(3 posts)→

Original post →

More from Research

Research channel →