The Origins and History of Machine Learning

September 29, 2018

Before this article gets on to the details of machine learning, it will enlighten the reader about the definition and concepts of machine learning. Some of the researchers and studies suggest that machine learning is a technique of artificial intelligence where the system automatically learns from information and data.

IT professionals define machine learning as a process by which the machine automatically improves their performance and algorithms without getting programmed manually.  Thus it can be said that machine learning history has several benefits like publishing sports reports, driving cars automatically and improved interaction with humans with the help of algorithms.

Recent surveys have suggested that examining the machine learning history will give the industries overview that how machine learning works and how it is important to the evolution on a daily basis. The collection from the data presented below will make the reader clearly understand the origins of machine learning and how it has evolved with the present date.


Turning test (1950)

This is a test which named after Alan Turing. According to this test, he was in the constant search of stating the machine has its own intelligence. Thus in order to pass this test, the computer must able to convince that it is also a human operating system and operates with its own set of intelligence.

Computer learning program (1957)

Computer learning program is basically the first ever computer program written by Arthur Samuel. The program was designed in the form of a game known as checkers. The system would play this game and through its course would develop winning strategies of its own to win the game. This would prove that just like humans the machine also had artificial intelligence to make changes according to its own convenience.

Neutral network (1950)

The neutral network was first designed by Frank Rosenblatt. The neutral network came to be known as a perceptron. Now basically this was a network which works on human brains, meaning using this network stimulation of human thought can be analysed and processed.

Nearest Neighbor (1967)

Nearest Neighbor is basically a creative algorithm which was designed in order for the system to evaluate or recognize the pattern. For example, the pattern could involve a salesman’s route to a particular city ensuring that the salesman covers all the major cities that fall within the route

Stanford Cart (1979)

Stanford cart is basically a program which was designed by the students of Ford University to find out the obstacles in a living room. With the help of this program, the system would automatically find out the obstacles in the living room using its own artificial intelligence.

EBL (1981)

Earlier till the 70’s, machine learning was all about programming but 1981 saw a major change when Gerald Dejong introduced the world to explanation-based learning. Now explanation based learning or better known as EBL was training data analyses which would allow the system to capture the important documents and exclude the unimportant ones.

Net Talk (1985)

In 1985 Terry Sejnowski developed a program known as the Net talk. With the help of this program, the system would be able to pronounce words just as a baby does. The system will take note of the pronunciation of the baby and will automatically generate sound and words that would be an exact copy of the baby.

Data-driven approach (1990’s)

During the times of the 90’s, there came a time when machine learning moved from a knowledge-based approach to the data-based approach. With the help of this approach, the IT professionals could evaluate large chunks of data and draw conclusions from it.

ERA of 2000

The era of 2000 includes some major changes that took place in 2006. 2011, 2012, 2014. 2015 and 2016. As early as 2006 developed a program known as deep learning which would allow the system to analyze the videos, images, and text on its own. Come 2010, Microsoft was able to track human features giving the humans the ability to interact with the system in the form of movements and gestures.

By 2011 Google was able to identify the objects just as the cat does and by 2012 Google developed a program which shows specific results when a particular search was made in YouTube. For example, if the user searched for the dog then with the help of these programs only videos of dog will be shown. This program provided as a blessing to précised searching. In 2015 machine came in the form of machine learning online. Both Amazon and Microsoft were able to launch their own machine learning platform. Now with the help of machine learning online data could be available on multiple machines and now better prediction could be made future made which was possible due to the batch learning process.


Hence to conclude it can be said machine learning has faced various transformations over the years. Now it can be said that the computer basically have their own artificial intelligence where it can think and act on their own. This statement is contradicted by various researchers and scientist who say that a computer will think the same as a human brain does. This is like comparing apples with oranges. Though it cannot be unnoticed that transformation of machine learning growing at random space and there is no limitation to it. The main question arising here is that will the computers be able to grow continually as the data keeps getting larger and larger.

Post a comment

nineteen + two =