-
McCulloch-Pitts neuron
Warren McCulloch and Walter Pitts made a paper on how neurons might work, and created a very basic neural network, called the McCulloch-Pitts neuron -
The Organization of Behavior
The Organization of Behavior is written by Donald Hebb, reinforcing the theory proposed by McCulloch and Pitts -
Period: to
the beginning of research groups
Nathaniel Rochester organized a group dedicated to attempting to create a neural network, whose experiments failed at first but succeeded in later attempts. Later on, another group known as the Dartmouth summer research project also began to study the concept of neural networking -
the perceptron
Frank Rossenblat created the idea of the perceptron, the first proper neural network. all modern day networks are based off this concept -
MADALINE and ADALINE
MADALINE and ADALINE, the first neural network used in real life applications, was created by Benard Widrow and Marcian Hoff, used to reduce noise in phone lines -
Period: to
Ingo's group
a group headed by Ingo Rechenberg was able to use genetic algorithms to solve complex problems in engineering -
the first of deep learning
1967: the first working deep learning algorithm was developed by Alexy Ivakhnenko -
Period: to
the ai winter
due to disappointment in natural networks (caused mostly by limitations of electronics), the fear of thinking machines starting to rise, and the publication of “perceptrons” by Marvin Minsky pointing out issues with, a large amount of funding was cut, and growth stunted. This was known as the ai winter -
the end of the winter
Growth started again due to a paper by John Hopefield and the us-japan joint conference on cooperative/competitive neural networks, where japan announced its fifth generation on neural networks -
the first NEAT
Ken Stanley develops the first NEAT (neural evolution of augmented topographies) artificial intelligence at university of Texas Austin
