History Of The Machine Learning Free Essay Example

To install StudyMoose App tap and then “Add to Home Screen”

Overview

Machine Learning is a current area of AI is centered on the ideas that we should just be able to feed machines with data and allow them to learn themselves from the data and improve their performance. AI has been around for decades but did not find use until now due to lack of powerful and sophisticated machine that we have and the absence of digital information invoked now. The Greek myths tells the stories of machines built to imitate human behavior.

Don't use plagiarized sources. Get your custom essay on

“ History Of The Machine Learning ”

Get custom paper

NEW! smart matching with writer

Early, European computers were thought of as “logical machines” and by possessing ability and capacities such as basic arithmetic and memory, engineers saw their work fundamentally as an attempt to build mechanical brains. Machine learning has emerged as a means of promoting the development of artificial intelligence at the current rate with two innovations. One of them was Arthur Samuel's 1959 achievement which states that rather than teaching computers everything they need to know about the world and how to carry out tasks, it might be possible to program them to learn for themselves.

The second is the recent rise of the Internet and the huge increase in the generation, storage and analysis of digital information. With these innovations in place, engineers understood that instead of teaching computers and machines how to do everything, it would be more efficient for them to think like a person and encode them to connect to the Internet for their own purposes to have access to all information in the world.

History

In 1642, a mechanical adder with wheels and gear was created.

One of the first mechanical adding machine was designed by Blaise Pascal. He used gear and wheel systems found in odometers and other computer devices. Well, consider what mechanical add-ons are in the history of machine learning. But if you take a closer look, you'll see that this is the first human effort to automate data processing. Pascal has developed a calculator to help his father make the arithmetic calculations that Luen had to perform as tax director. The machine was built to sum and find the difference between two numbers, and to carryout multiplication and division by repetitive addition and subtraction respectively. The adding machine has a metal wheel dial which has figures from 0 to 9 around each wheel. To enter a figure, the user places the stylus in the space between the spokes and the dials until it reaches the lower metal stop similar to the way the rotary dial of the phone is used. This will show a number in the window at the top of the calculator. Then, recompose the second number to add to display the sum of the two numbers in the accumulator. One of the most striking features was a porting mechanism that adds 1 to 9 to a dial, and when you go from 9 to 0, it goes 1 to the next dial.

Data storage was the next challenge in 1801. The first use for storing data is to punch holes in a card invented by Joseph Marie Jacquard. These cards were used to code the program that shows the loom which allowed the process to be iterated with consistent output each single time. The Jacquard loom automatically obtained the desired pattern using an exchangeable punch card that controls the weaving of the fabric. The famous British inventor Charles Babbage adopted the punch card as an input and output medium for the analysis engine he proposed and used by the United State statistician Herman Hollerith to serve as an input medium to his census machines. They have also been used as a means of data input on digital computers, but replaced by electronic devices.

Boolean logic which is a way of making arguments or inferring them as true or false conclusions was created in 1847. George Boole created a way of response using boolean operators (AND, OR, NOR). The answers are displayed as true or false, or as yes or no, and as 1 or 0 in binary. Web search always uses this operator. Herman Hollerith built the first statistical machine using punch cards to process data received from millions of people. The machine was built to summarize the data on the punch card. Also in 1890 aggregation devices which its latest models have been widely used in commercial applications such as accounting and inventory management were developed to help process US census data.

Alan Turing, an English mathematician, pioneer of artificial created in 1950 the "Turing Test" to determine if a computer had real intelligence. To pass the test, you must believe that your computer can fool people. According to this test, a computer is considered to have artificial intelligence if it can respond like a human under certain conditions. Three main points exist in the turing test. Two are controlled by the human operator and the last point is controlled by the computer. Each point is physically different and distant from the other two points. The computer is questioned by a human and the computer is expected to generate responses within a time interval. The interrogator will analyze how the computer responds and when the computer is running by testing it repeatedly. Computers are considered endowed with artificial intelligence, if the computer makes correct responses and the interrogator cannot tell if it was a human or computer responding at the other end. Arthur Samuel wrote his first computer learning program in 1952. The program was an improved computer game designed to move from a bet to a studio, a chess game and other IBM games to learn cards in mode and integrate these movements into the program.

Frank Rosenblatt in 1957 came up with perceptron, a network of neurons modelled like the brain. The brain contains billions of cells called neurons linked to one other in a network. Perceptron links the web to the state where the easiest decisions are gotten from bigger programs to solve complex issues. In 1967, "nearest neighbors" algorithm was created to allow computers to start a simple recognition. When a new object is given to a program, it is classified as the nearest neighbor, which is the closest object in memory, compared to the existing data. This can be used to map routes for travel vendors from any city. It allows you to visit all the cities for short trips. A Stanford university students invent the "Stanford Cart" in 1979, which allows students to explore obstacles in their rooms. Stanford Cart was a mobile robot equipped with a remote control. A computer program was created to guide the cart in a congested space and gain a knowledge of the world with images transmitted by the on-board television system. The basket looked in three dimensions for a three-dimensional object using a multidimensional stereoscopic vision and deduced its movement. Based on the model built with this information, a route to avoid obstacles to the desired destination was provided.

Gerald Dejong published an article on a learning based description of the world's prior knowledge in a newspaper in 1981, which is provided by an example of supervised learning. Given guidelines for achieving goals, the program will analyze the training data and reject the irrelevancies to form the general guidelines to follow. For example, if an algorithm allows chess to focus on the queen, it refuses to accept all parts that have no immediate effect. In the nineties, machine learning applications began in data mining, adaptive software and web applications, text learning, and language learning. Scientists have started to create computer programs to analyze large amounts of data, draw conclusions, or learn results. Machine learning has been able to develop a program that allows not only the advancement of technology, but also the defined once-learned things to be learned and to continue to evolve by introducing new data without the intervention of those who need it. An achievement in machine learning in 1997 was Deep Blue of IBM which defeated world chess champion and the new millennium in 2000 which brought development in adaptive programming. The program recognizes patterns, learns from experience, and constantly improves with feedback from the surroundings. One example of adaptive programming is in-depth learning where algorithms can "see" and distinguish between objects in images and video.

IBM Watson 2011 is superior to human competitors in Jeopardy. 2011 Google GOOGL + 0.52% brain has been developed and the neural network can learn to discover and classify objects in the same way as cats. 2012 Google X Lab develops machine learning algorithms that can independently navigate YouTube videos to identify videos that contain cats. Facebook in 2014 came up with Deepface, a system that recognizes individuals in digital images. Amazon also started its own machine learning platform likewise Microsoft which created the Distributed Machine Learning Toolkit, which enhances effective deployment of machine learning on multiple computers. Artificial Intelligence and Robotics, supported by Stephen Hawking, Elon Musk and Steve Wozniak (among others), more than 3000 researchers in 2015 open letter of warning about the risks of participating self-defense weapons without the choice and human intervention. The Google Artificial Intelligence 2016 algorithm is considered to be the most complex board game in the world and better than the Chinese board game athlete which is much more difficult than chess. The AlphaGo algorithm, developed by Google DeepMind, has led 5 games out of five in the Go contest.