Artificial Intelligence
What is artificial intelligence all about anyway?
A brief summary of Artificial Intelligence
Artificial intelligence (AI) is the future. Artificial intelligence only belongs in the movies. Artificial intelligence is already part of your everyday life. All of these are true in their own right, it depends on what kind of AI you are referring to.
If you think AI is a thing of the future, I’d like to give you this example to show how AI has been used for many years. We all rely on apps to get us around, a long time has passed since the average person has relied on a physical map to navigate through cities, countries and continents. One app that is extremely popular is Google Maps. Google uses your location data from smartphones. and can analyse the speed and direction of movement. It can show you the direction to go with live traffic updates at any given time.
Data is key to AI and in the future data will be the new currency of governments, businesses, and organisations. Without data AI would be impossible …In the case of Google Maps, data is collected and AI algorithms are trained to improve your daily commute by reducing time.
In the media they may use terms such as artificial intelligence, deep learning and machine learning interchangeably to describe these technological advancements, but they are not exactly the same thing, I will explain the difference in another blog but in general the key differences are shown in the diagram below.
Artificial intelligence is the larger circle that encapsulates Machine and Deep Learning. AI is classified as an intelligent machine that leads to an optimal or sub-optimal solution given a specific problem.
Wikipedia gives the following definition of an Artificial Intelligence program, as any device that perceives its environment and takes actions that maximize its chance of successfully achieving its goals.
Programs that fall under the canopy of AI but not machine learning tend to be programs that can utilize decision trees for logic, or programs that build rules and instructions for potential problems that may arise.
For example, if you make a purchase on a credit card in a ‘newly visited’ or ‘far from home’ country, then your bank can and should flag all of its transactions for review. Some banks may even freeze your card completely until you confirm you have used your card in another country (good for fraud but when you’re travelling it can be the bain of your life… as you constantly have to update your bank as and when you will visit each country). There may be no training data for this, so there will be no machine learning; it’s a simple rule encoded by a programmer.
On the other hand, other things might be handled by ML: for example the number of digits in an email address, combined with the domain of the email, combined with the IP network, combined with the time of day for the transaction, might all be features which push a ML system over the edge to reject a transaction online.
As you can see from the diagram, AI is a broad concept that was founded in the 1950’s. AI can be further generalised into categories based on AI’s ability. These are Artificial Narrow Intelligence, Artificial General Intelligence and Artificial Super-intelligence.
Artificial Narrow Intelligence specialises in one area. For example there is AI that identifies your face on images on Facebook, but that’s the only thing it does. If you were to ask it to drive a car, it wouldn’t know where to begin.
Artificial General Intelligence refers to a computer being as smart as a human across a range of areas. General intelligence would mean a machine could perform any intellectual task that a human can. Accomplishing this is a LOT harder than Artificial narrow intelligence, and has not been done yet, so we are along way from iRobot don’t worry! A machine would have to have human qualities such as be able to plan, solve problems, learn quick, learn from experiences, one-shot learning, abstract thinking and many more to have general intelligence.
Artificial Super-intelligence defined by AI thinker and Oxford philosopher is “an intellect that is much smarter than the best human brains in every field, including scientific creativity, wisdom and social skills.” Artificial super-intelligence would range from computers a little smarter than humans to one that’s exponentially smarter.
Artificial narrow intelligence already exists and is everywhere in daily life. Just like the industrial revolution, many people refer to AI as starting an intelligence revolution. Where an AI will transform from narrow into general AI and then exponentially improve in levels of intelligence to super -intelligence.