Deep Learning - IIT Ropar

From biological neurons to perceptrons

This NPTEL course from IIT Ropar, taught by Prof. Sudarshan Iyengar, runs to 117 lectures. It opens with the biological neuron and the history of AI, then builds through McCulloch-Pitts neurons and thresholding logic to perceptrons and the perceptron learning algorithm.

Later topics

• The deep revival and the path from early image work to convolutional networks
• Sequences and why they call for their own architectures
• A proof of convergence for the perceptron learning algorithm

Who it suits

Learners who want the history and building blocks of neural networks at lecture-length detail will find it a solid start. The early lectures lean on the perceptron, and the convergence proof is a good example of the mathematical reasoning the course uses.