Lesson 16 of 21 · Part V, Deep learning and beyond

Convolutional neural networks (CNN)

Local connections, shared weights, pooling and many layers: 1D and 2D convolution, the receptive field that grows with depth, LeCun’s networks, MNIST, AlexNet and ImageNet.

13 figures to manipulate.

In this lesson

• The problem: recognizing characters
• Convolution
• A network of “filters”
• Pooling
• The CNN as a whole
• How they are used
• Historical results
• Modern CNNs and GPUs

Other lessons

Bias-variance and ensembles · Deep learning · All lessons

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