Warren McCulloch and Walter Pitts were the first to create a model of artificial neural networks back in 1943. They built the model on something called threshold logic. A threshold was calculated by summing up inputs, and the output was binary, zero, or one, according to the threshold. In 1958, another model of a neuron was created by Rosenblatt called perceptron. Perceptron is the simplest model of an artificial neuron that can classify inputs into two classes (we discussed this neuron in Chapter 1, Getting started with Deep Learning). The concept of training neural networks by backpropagating errors using chain rule was developed by Henry J. Kelley around the early 1960s. However, backpropagation as an algorithm was unstructured and the perceptron model failed to solve that famous XOR problem. In 1986, Geoff Hinton, David ...
Germany
Slovakia
Canada
Brazil
Singapore
Hungary
Philippines
Mexico
Thailand
Ukraine
Luxembourg
Estonia
Lithuania
Norway
Chile
United States
Great Britain
India
Spain
South Korea
Ecuador
Colombia
Taiwan
Switzerland
Indonesia
Cyprus
Denmark
Finland
Poland
Malta
Czechia
New Zealand
Austria
Turkey
France
Sweden
Italy
Egypt
Belgium
Portugal
Slovenia
Ireland
Romania
Greece
Argentina
Malaysia
South Africa
Netherlands
Bulgaria
Latvia
Australia
Japan
Russia