Machine learning dates back to the 17th century. Machines can learn tasks not explicitly programmed for. We can now train machines to operate in unknow environments. What is the history of machine-learning? Continue reading to learn more. This is an important topic for engineers as well as computer scientists. If you're interested in the history of machine learning, you'll appreciate this article.

Neural networks

Walter Pitts McCulloch and Warren Sturgis McCulloch were the first to invent artificial neural network technology. Their work was essential in establishing the foundation of neural networks. The two scientists demonstrated that an output can only be active if an input is active using logic gates. The two scientists were able simplify the functioning the brain and opened the door to machine learning.

Convolutional neural networks

Convolutional neural systems are made up multiple layers of artificial cells. Each neuron of the network is a mathematical operation that calculates a weighted combination of its inputs, outputs, and a value called activation. When they are given pixel values, artificial neurons learn to recognize visual features. CNN has a first layer of convolutional layers. This layer contains the input image and generates activation maps. These maps highlight different parts of the image.

Boosting

Although the concept of boosting isn't new, it was first used in machine learning in the 1990s. It is an algorithm to improve supervised learning by turning weak learners into more competent ones. Robert Schapire was the first to propose the concept of boosting. In 1990, he wrote a paper that described how weak classifiers could be made strong. Weak learners have a low correlation with the true classification. On the other hand, strong learners are closely aligned to the real classification.

Turing test

The Turing Test has become one of the most important concepts in the philosophy of artificial intelligence. Using a computer as the subject of an interrogation, the machine must be able to produce an enquiry that the human interrogator does not understand. Turing Test passes machines that are capable of doing this. The problem with this test is that it attracts projects whose primary aim is to fool the judges.

Deep learning

Machine learning and deep learning have a long history. It all started in 1965 when Valentin Grigoryevich Lapa and Alex Grigoryevich Ivakhnenko developed an algorithm that utilized polynomial activation function. The idea behind the algorithm was to recreate the brain’s neural networks using data analysis. Although funding was unavailable for research into artificial Intelligence in the 1960s, individuals continued to pursue the topic.