Lesson 6 of 21 · Part II, Linear models and neural networks

Neural networks (part 1) — from the neuron to the MLP

From the biological neuron to the Perceptron and its convergence theorem, sigmoids, the Multi-Layer Perceptron as an adaptive basis expansion and universal approximator.

15 figures to manipulate.

In this lesson

• Neural networks as an ML tool
• The biological inspiration
• The artificial neuron
• The Perceptron
• Learning for a single unit
• Perceptron versus LMS
• Sigmoidal activation functions
• Neural networks: the Multi-Layer Perceptron
• Why neural networks are flexible
• Toward the learning algorithm

Other lessons

Linear models and K-nearest neighbors · Notes on backpropagation · All lessons

Notes by Fabio Piscitelli — Machine Learning (654AA), Prof. Alessio Micheli, University of Pisa, a.y. 2026/27.