Intelligence is not Artificial

by piero scaruffi

Cognitive Science and Artificial Intelligence | Bibliography and book reviews | My book on consciousness | Contact/feedback/email
(Copyright © 2018 Piero Scaruffi | Terms of use )


(These are excerpts from my book "Intelligence is not Artificial")

Trivia: What is the Homeland of Deep Learning?

Fukushima is Japanese, LeCun, Bengio and Collobert are French, Hinton and Zisserman are British, Thrun, Schmidhuber, Hochreiter, Szegedyand Socher are German, Ng, Fei-fei Li, Yangqing Jia, Aja Huang, Kaiming He and Dong Yu are Chinese, Krizhevsky and Sutskever are Russian, Olshausen is Swiss, Goodfellow is Canadian. Add Siegelmann from Israel, Daniela Rus from Romania, Quoc Le from Vietnam, Malik from India, Simonyan from Armenia, Karpathy from Slovakia, Abbeel from Belgium, Vinyals from Spain, and the DeepMind founders from Britain (Demis Hassabis, son of Greek man and a Chinese woman, and Mustafa Suleyman, son of a Syrian man and a British woman) and from New Zealand (Shane Legg, a student of Schmidhuber in Switzerland).

What do they have in common? None of them was born in the USA. Many of them now work in the USA or for US companies, but they were not born there and many were not educated there.

There was a period of time during which US universities (especially the heavy weights Stanford, MIT, Yale and Carnegie Mellon) shunned neural networks, and US-born students had little motivation to work on such an esoteric and useless subject. The flame of neural networks was kept alive in the countries where symbolic A.I. did not completely obscure neural networks, which is pretty much every country except the USA.

Japan has always had a strange attitude towards software and neural networks were probably more appealing in that country for being so simple to program (and so difficult to implement on the hardware).

In 1982 Canada opened the Canadian Institute for Advanced Research (CIFAR), a center that was independent from the schools of thought in the USA, and whose founder, James Mustard, was certainly closer to brain studies than to software. Europeans mathematicians were attracted by the mathematics of neural networks. Luckily, in 1988 the Italian philanthropist Angelo Dalle Molle established in Lugano (Switzerland) the Istituto Dalle Molle di Studi sull'Intelligenza Artificiale, another center insulated from academic pressures.

In 1991 British physicist John Taylor (who had once delivered a lecture titled "A Model of Thinking Neural Networks", 1973, and who dabbled in paranormal phenomena) founded the International Conference on Artificial Neural Networks (ICANN), the first edition being held in Finland.

Whatever the reason, deep learning was the outcome of a truly distributed and international collaboration from which the USA was actually notably missing.

Back to the Table of Contents


Purchase "Intelligence is not Artificial")
Back to Cognitive Science | My book on consciousness | My reviews of books | Contact