You might have seen Iron man’s “Jarvis” or “Vision” from Avengers or “The Terminator” in Hollywood movies.These are the example of Artificial intelligence in cinematic world.
It’s very interesting to know, how they work. They are not human but a human developed computer program which can listen, think and react.
“Sophia” : The social humanoid robot is the real world example of Artificial intelligence. She became first ever robot to get citizenship (Saudi Arabia) of any country in world.
Do you know?
In 1996: IBM’s Deep Blue became the first machine to win a chess game against a reigning world champion, Garry Kasparov.
In 2011: IBM’s Watson beat Jeopardy champion Ken Jennings in a question-answer american TV show “Jeopardy” (KBC of India).
In 2016: Google’s AlphaGo defeated legendary 18 times world champion Go player Lee Sedol by 4-1 and Lee left the game before 28 minute and 28 second by seeing Alpha Go’s perfect move.
These are the artificial intelligent supercomputers or robots.
Beating in chess game or solving question-answer round might less impress you but Go is a different game. It can not be predicted. Over 10 to 170 moves are possible in Go.
Alpha Go was trained by real human Go games. It ran through million of games and learned the technique used and created some new one that no one had ever seen.
Very few people know that only a year ago after Alpha Go’s Victory, A brand new AI Alpha Go Zero beat original Alpha Go with 100-0. The interesting thing about Alpha Go Zero, It learned to play Go without any historical data or figure given and without any human intervention. With just bare bones rule, Alpha Go Zero surpassed previous Alpha Go and 2500 years of strategy and knowledge in just 40 days of learning. It only played against itself and now it is regarded as best go player in world even though it is not human.
So how these robots did it or what is artificial intelligence? Lets understand about this.
What is Artificial intelligence?
The word Artificial intelligence first coined in 1956 by John McCarthy at Dartmouth conference, MIT.
Artificial = Non-Natural (Made by human) and Intelligence = Ability to understand, think and learn.
Artificial intelligence(AI) is a form of computer science which is used to create intelligent machine which can make decision by their own, can recognize human speech
or objects, can learn, plan and solve problems like humans.
How Artificial intelligence work?
Machine learning is the backbone of AI. Machine learn by itself. To understand the machine learning, we need to first understand the human brain.
Human brain is made up of millions of neurons which communicate with each other and process the information in brain and make us intelligent. For example, If a person having pulse and breathe both, it will tell that person is alive and brain will send the signals to the body. This world is full of sounds, images ( Data in general). It will take all of them to form our view of reality.
The same thing we achieve through the machines. However, when more complex data is shown, it become harder to analyse. That is where machine learning came handy.
This below image shows a picture of human neuron and computer neuron. Function of x works as neuron system for machines which take multiple x as an input and process output.
Machine can not only analyze the data but also learn, adapt or view from it. Alpha Go Zero is the example of this which surpasses 2500 years of strategy and knowledge and even made its own new strategy in just 40 days.
But Is this question coming in your mind ,that how did it do this so much faster and with the efficiency?
Generally, Biological Neural network operate around 200 hertz and a transistor operate over 2 Gigahertz. The neuron impulses in brain traveled through axon about 100 meter per second which is pretty fast and give us a good reaction time but its only third as fast as the speed of sound.
Computer however can transmit the information at speed of light or 300 million meter per second. So there is quite big difference between our human brain capability and a computer. In just one week, computer can do 20,000 years worth of human level research or simulation that is trained to do so.
Normally, we give data and algorithm to take out result from computer but In Machine learning, the data and desired output is provided to machine to create the learning algorithm.
You must have listen, Practice makes a man perfect. Similarly, Machine keeps on reading or practicing with the data and desired result to improve its efficiency and accuracy.
As an example, For recognize a dog picture, thousands of dog’s picture will be feed in computer. Now it will create some parameter and compare a new dog picture with those parameters. It is fits accordingly, this will tell that it is a dog and it will feed this picture as well to compare next new pictures.
Every dog has different structure of face. Machine will read every face and records its parameter to learn and make itself accurate to recognize when new face is shown.
Apple’s face recognition system is an example of this.
Image processing and Neural network is the fundamental of learning algorithm. Speech recognition or image recognition is possible just because of learning algorithm.
Machine learning and AI are same?
Machine learning is just a part of AI. Using machine learning , A machine can do a specific thing based on past data But AI can learn new thing and do multiple
For example, A Chess playing machine will play only chess. It will fail to play another sports like cricket. But AI will try to learn other games when it will come in
contact and In some time, it will become accurate to play both chess and cricket.
In short, Machine learning work on past experience but AI work on building future and enhance itself.
How to classify AI?
Generally we classify the AI in 2 types : Weak AI and Strong AI
These terms have nothing to do with the physically Weak and strong. There are some difference between these two which is explained further:
Weak AI : This AI is only exist as of now. It has been used in many machines and further improvement is going on. This AI is dependent on human inputs to improve its results. Example are self driving car, computer games or google translator.
Self driving car simulate the human intelligence by taking input through surrounding and past outcomes and move ahead. In computer game, sometime computer defeat us because it is not the computer playing against us. Actually, It is the tactics which is added by another person in the algorithm previously. Google translator also keeps on improving itself by taking different kind of inputs from user. There are many other examples for this kind of AI in our life.
Strong AI : This AI is where we are trying to get to but it is not sure whether this will be possible in future. This AI has a human level intelligence and perform task without any external help.
Cinematic character Jarvis, Vision and Terminator are the example of this. In this AI, Machine learn, think and perform intellectual task like humans. Human have emotions,consciousness and he can innovate anything. This is difficult to teach the machines. This is called as artificial general intelligence. These AI machine can create the computer more powerful than itself.
Application of AI :
AI has infinite possibilities to be used for. In normal days, we use the AI without even realize it. When we ask something to SIRI in iPhone or say “OK Google” to google assistance, there is an AI which is working behind for us. When we search something on Google, some kind of AI is getting us the perfect results.
Amazon echo, Google home, Self-driving cars, Facebook face recognition, Amazon recommended products are another example of AI. We can use AI in banking for fraud detection, online virtual assistance ( Ira and Eva of HDFC, Siri, Cortana), Cyber security to detect and respond to threats, home robots to help people and many more places.
Huawei CPU Kirin 970 AI processor has builtin neuro-processing unit. While clicking photo, It can process 2000 images/sec and automatically change setting according to
background and you will get a good picture.
What is the Future look like of Artificial intelligence?
Once Tech entrepreneur Elon musk said about Artificial intelligence that It could be terrible and it could be great, Its not clear but one thing is sure, we will not control it.
When this question was asked to humanoid robot Sophia in an interview, She replied below:
Facebook founder Mark Zuckerberg told Musk’s prediction as “Imaginary Fear”. Swedish philosopher Nick Bostrom said, If AI is not managed carefully, humanity is engineering risks to its own extinction.
Free courses for artificial intelligence:
There are several paid course in market which charge you high but Google provide a free learning platform if you are interested in AI.
Well, We don’t know what direction AI will take in future but human will be able to augment them self with robots. Robots will take hazardous jobs i.e. Bomb defusing, welding etc. Home robots will help people. Your smartphone will become your friend. It will talk to you and detect your emotions and will suggest you some ideas accordingly for read or travel to make your day good.
You would also like to read: