Artificial Intelligence is becoming quite popular these days, not just in the computing world but in general public imagination too. Thanks to Hollywood movies like The Terminator, Artificial Intelligence has been associated with a lot of things.
So, what is Artificial Intelligence actually? Read the following sections that highlight some basics about AI and how AI works.
What is Artificial Intelligence?
Before defining Artificial Intelligence, take a look at the following examples:
- A computer app that plays chess
- A self-driving car
- Alexa responding to your requests
These are few examples that make use of AI at a very basic level.
A computerized chess game contains a database of chess moves and works by finding the right move from its repository in a given situation.
A self-driving car uses imaging software and sensors to drive on its own without any human intervention.
Alexa stores your preferences and uses them to “guide” you on what you might want to watch or listen next.
If you look perfunctorily then you will think that the above examples resemble human actions. A computer is thinking like a human while playing chess or Alexa is helping you make the right choices just as a friend of yours would do.
The above computer programs are making decisions, a very human like behavior.
And that ability to make decisions is just one goal of artificial intelligence.
AI aims to make machines that can reason, plan, decide, learn and speak in ways similar to humans.
And, that is what Artificial Intelligence is actually all about. It is the branch of science that uses statistics, analytics, physics, voice simulation, imaging and much more to come up with a program that thinks and acts like humans in a given situation.
Arnold Schwarzenegger in and as Terminator, is a very far, far-fetched idea of what AI is or can be!
AI is only making machines work smartly. These machines are still dependent upon data input by humans. Also, they are in no way close to human intelligence. For example, a machine programmed to play chess will not be able to play any other game or drive cars. Today, the state of AI is – one machine, one task.
How Artificial Intelligence Works?
AI has many subsets that together help any robot or a computer program to think and act like a human would in a specific situation.
The subsets of Artificial Intelligence are:
Machine learning as the name implies is the knowledge gained by the machine on its own. The machine looks at the given data and the past data, observes patterns and makes informed decisions by itself.
The machine has not been programmed to act in this way instead the machine learns by itself and makes decisions.
Deep learning is a type of machine learning that requires huge amount of data in the form of neural networks. It also relies on high processing power to come up with outputs.
Self-driving cars, image recognition software and speech processing software use deep learning.
The concept of neural network is based upon the working of neurons in the human brain. Like neurons in the brain transmit information from one part of the brain to another, similarly neural networks in artificial intelligence process and pass information from one network to another.
Natural Language Processing and Speech Recognition
Natural language processing uses languages like English to communicate with humans. It stores and processes the information and gives the output which is in human understandable format. The format can be in the form of speech or text.
Some examples of AI that use NLP are Siri, Alexa and Cortana.
Robotics involves the creation of robots that can work and interact with humans. The robots are programmed in a way so that they can respond intelligently.
The latest use of robotics in AI is Robot Sophia.
The aim of machine vision is to provide robots/computer programs the ability to see and recognize objects.
As you can see the different subsets of AI are trying to come up with workable models of machines that mimic the various activities of human brain, like vision, speech, memory, learning and so on.
Each subset is either working independently or in tandem with other subsets.
Artificial Intelligence is of two types: weak/basic and strong/advanced. The examples of AI machines that you see today are of weak or basic AI types. A strong or advanced AI machine has not been made yet.
What AI has been able to achieve till date is very basic, weak and narrow form of artificial intelligence.
So, will AI ever be able to create a Terminator type of robot? Scientists are not very sure on this. Because human mind is very complex, duplicating it looks almost impossible.
- Medical science – Robots that remind you of your medicines, sensors that can determine whether a particular cancer is benign or malignant.
- Finance sector – Computer programs that help detect frauds and aid in automation.
- Retail sector – Helps a consumer make purchase decision.
- Entertainment sector – Helps choose programs/shows based upon a user’s past choices.
- Automotive sector – Self driven cars.
- Gaming – Programs that can play Chess, Go.
Too much reliability on machines can be dangerous especially in areas like medical science or financial investments. After all machines work on data that has been given by humans. Many times scientists themselves are unable to explain the inner workings of a machine in decision making, which makes it difficult for layman to blindly trust the judgment of machines.
Wrap up – Artificial Intelligence and How it Works
Human mind is versatile and it is almost impossible to make a copy of it. This is the reason Artificial Intelligence is still in nascent stages and it might take years to come up with a machine that even remotely thinks like humans.
Till then, just relax and enjoy the little marvels that AI keeps churning up to make life somewhat easier.