Franco Raimondi – Perceptrons at the helm

“We invented a thinking machine” Herbert Simon, 1956

The term Artificial Intelligence (AI) was first used by John McCarthy between 1955 and 1956 to refer to a discipline aimed at building systems that could simulate intelligence, however the attempts to create intelligent machines had been underway for decades. The mathematical and technical foundations for the development of autopilots were defined around 1920, and cybernetics (from kybernétes, meaning helmsman) emerged in the 1940s. In 1943, Warren McCulloch and Walter Pitts proposed the first mathematical formalisation of a network of artifi cial neurons. This was implemented on a computer in 1957 by Frank Rosenblatt under the name Perceptron, and it remains today the basic component of Neural Networks. Meanwhile, fears associated with the loss of control over automatons also developed: the term ‘robot’ was introduced in 1920 in the dystopian play R.U.R. (Rossum’s Universal Robots), which ends with the extermination of humanity. Seventy years after the birth of the term Artificial Intelligence and one hundred years after the publication of R.U.R., are we really at a turning point?

This book provides a brief historical introduction and clearly presents the concepts of Artificial Intelligence behind the applications develo ped in recent years.