This page contains general information regarding Artificial Intelligence. Although the main website focuses on machine learning, it is helpful to have some background knowledge regarding where machine learning originated from.
What is Artificial Intelligence (AI)?
AI is an ability of a machine or computer to learn and make decisions similar to those humans make. This ability varies by machine, but ultimately gets at everything people can do, from face and speech recognition to playing chess.
Main Branches
There are many sub-fields to AI, but most of them fall within one of these two categories.
Search and Optimization: This branch gets at everything from planning (like in Google maps) to function optimization (finding the best algorithm to solve a situation).
Machine Learning: Machine learning involves having the machine learn something from facts or experiences as opposed to spending the time programming it. It is often used particularly in situations that involve Big Data ("data sets that are large, diverse, and fast changing" (Hack)). Since each time the machine analyzes a situation or performs a task it improves, machine learning results in machines that are more accurate and faster, making them more useful when working with Big Data. The machine learns from not only the information given to it but also from its past experiences. Some common examples are pattern recognition, Clever-bot, and Siri. More details on some other examples are on the rest of the website.
Search and Optimization: This branch gets at everything from planning (like in Google maps) to function optimization (finding the best algorithm to solve a situation).
- Search: Search problems are basically path finding problems that involve finding the steps and routes to a solution. The methods used could be through algorithms or a brute force approach. For example, looking up directions on your phone would be a path finding problem.
- Optimization: Optimization problems are what they sound like. Given a certain problem and variables in the problem, find the most efficient, least resource intensive solution to it. For example, plot the route from your house to a school. Google maps would accomplish this as an optimization problem and give you the fastest route with the least traffic.
Machine Learning: Machine learning involves having the machine learn something from facts or experiences as opposed to spending the time programming it. It is often used particularly in situations that involve Big Data ("data sets that are large, diverse, and fast changing" (Hack)). Since each time the machine analyzes a situation or performs a task it improves, machine learning results in machines that are more accurate and faster, making them more useful when working with Big Data. The machine learns from not only the information given to it but also from its past experiences. Some common examples are pattern recognition, Clever-bot, and Siri. More details on some other examples are on the rest of the website.
Relationship Between Machine Learning and AI
As AI developed, machine learning developed alongside it as a branch focused on data interpretation. In the 1990s, as AI moved more toward trying to replicate a more human form of emotion and reasoning, machine learning diverged from AI. Then, beginning in the 2000s, AI and Machine learning gradually came back together, since data interpretation is recognized as an important element of simulating a thought process.
Important Events
1948: Norbert Wiener publishes “Cybernetics”, which was very influential in AI research.
1950: Alan Turing published an essay called “Computing Machinery and Intelligence”. He also made the famous Turing
test, in which if passed by a computer, it can be said to be as intelligent as a human.
1956: John McCarthy coined “artificial intelligence” in an AI conference.
1959: UNIMATE began working at General Motors as the first industrial robot.
1964: Danny Bobrow shows that computers can solve word problems.
1965: ANALOGY, a program written by Thomas Evan, demonstrates that computers can solve IQ test analogy
problems.
1970: Stanford Research Institute introduces Shakey, the first robot controlled by AI that had the capability of moving around.
1981: Danny Hillis creates the connection machine that brought new power to AI.
1990s: Machine learning started to move away from AI's symbolic methods and towards models from statistics and
probability theory.
2000s: Machine learning and AI started to come back together.
1950: Alan Turing published an essay called “Computing Machinery and Intelligence”. He also made the famous Turing
test, in which if passed by a computer, it can be said to be as intelligent as a human.
1956: John McCarthy coined “artificial intelligence” in an AI conference.
1959: UNIMATE began working at General Motors as the first industrial robot.
1964: Danny Bobrow shows that computers can solve word problems.
1965: ANALOGY, a program written by Thomas Evan, demonstrates that computers can solve IQ test analogy
problems.
1970: Stanford Research Institute introduces Shakey, the first robot controlled by AI that had the capability of moving around.
1981: Danny Hillis creates the connection machine that brought new power to AI.
1990s: Machine learning started to move away from AI's symbolic methods and towards models from statistics and
probability theory.
2000s: Machine learning and AI started to come back together.