The Perceptron as a Roadmap (Graduate Colloquium in Math, Philosophy and Physics)

Document Type

Conference Proceeding

Streaming Media

Publication Date



The main character of the lecture is the perceptron, which was created by the psychologist Frank Rosenblatt in the sixties of the previous century, to be an image classifier in the study of vision. The perceptron model was based on the 1943 model of the neuron due to McCullogh and Pitts. The perceptron algorithm is based on the structure of the neuron, and can be seen as one of the first, if not the first, machine learning algorithm. In the lecture, using the perceptron as a roadmap, we describe the evolution of the ideas from the discovery of the structure of the neuron to the structure of artificial neural networks and machine learning. Links with Hopfield networks and associative memories will be discussed as well as links with function approximation. Mathematicians in the audience might be surprised to see names such as Agmon, Schoenberg and Wiener pop up in the discussion.


Professor Daniel Alpay: "From the structure of the neuron to artificial neural networks: the perceptron as a roadmap"

A session of the MPP Seminar / Graduate Colloquium