An attempt to emulate the brain in image recognition in 1940.  The first perceptron was built in 1958.  A breakthrough occurred in 1988 with CNN (convolutional neural network).  CNN deals with spatial relationships.  Another breakthrough was the discovery of RNN (recurrent neural network) which deals with sequential relationships.  Machine learning is largely a black box and not completely interpretable.  Each learned model is highly related to the data on which it was trained, and the learning is not very transferrable to new data.  It needs a lot of data and computing power for training.

	CNN 1988