Friday, September 15, 2017

What is Artificial Intelligence?

Artificial Intelligence has created a lot of buzz around the world these days. You may have heard about the Elon Musk's Bot defeated the Champion gamer in dota2. Also, you have read Facebook closing its AI Company due to the Chat-Bots creating their own language and negotiating within themselves. And the funny thing is that two Chinese Chat-Bots created in China were abusing their own system. Also, now every smartphone has one or the other voice assistants embedded. Talk about Google Assistant, Bixby, Amazon echo, or Siri all assistants provide us with many smart features. Basically, now there are many machine hardware or computer program that are very smart and does a lot of intelligent stuff’s that were not possible few years before. 


What is Artificial Intelligence?
Putting simply, it is the ability of machines and software programs to be able to think and make decisions like humans. AI machines solve a problem like us humans without having Natural Intelligence. Machines will also improve itself while solving the problem and they are known as Self-learning AI’s. Definitely, AI is not only about thinking and taking decisions, it’s a lot more.

There are a few factors on which Intelligence is rated:-
1      Generalization Learning:- It is the ability of machines to perform better in a situation that is not previously encountered. For e.g. If a machine has seen a few different pictures of a dog previously, it should be able to identify a new different picture of a dog.
2      Reasoning:- A Machine should be able to draw some conclusions appropriately with the data it has read, given a particular situation in hand. For e.g. Think about a Doctor Bot who can recommend you medicines after you have mentioned your health issues.
3      Problem Solving:- It’s the ability of machines to solve some computational problems. For eg., you have given some data to a machine for finding the value of X.
4      Perceptions:- It’s the ability of machines to scan a particular environment and differentiate between Objects. For eg. Self-Driving Cars like Tesla, they have many sensors installed on them to differentiate between cars, trees, people and many other things on road.
5      Language Understanding:- It's basically to understand human language through syntax formation.

      Applications of Artificial Intelligence

Machine Learning:- It’s the ability of machines to learn things iteratively with being explicitly programmed. Basically, Machine’s learn from their previous experiences and perform better in situations that are not previously encountered.
Computer Vision:- It’s a Science that aims machines should visually sense the world around them. Basically, computers extract, analyze and they understand the information from visual pictures. 
Neural Network:- Neural Networks are simulated networks like human nervous system. The system uses nodes that act as artificial nodes and many nodes are connected together to connect different information so that machines can learn automatically by only observing things and don’t need coding. It is based on Input and output system. Take an ex. A Programmer doesn’t how to make the computer understand a picture of a dog by coding, but he can do it easily by neural networks, we just have to show computer different pictures of dogs so that it processes it and can connect how the dog looks like.
      Although we have learned about some of the basic applications of Artificial Intelligence, in short, we need to go through each of them in greater detail because each topic has got some of its own uniqueness. I will cover most of them in later posts.

   How can it be Destructive?

      I have heard that humans are the most ambitious species on earth. But, It’s not wrong if I say machines can be Over-Ambitious as they are made by humans itself. Just think if they become more intelligent than the most intelligent human on earth. They can definitely be misused. You can relate if you have seen Avengers or Robot. We also have some evidence from the past like Google’s deep mind challenge Match between 18-time World champion Lee Sedol and AlphaGo played in Seoul, South Korea in March 2016 where the Google’s Bot Alpha Go defeated all the matches to Lee Sedol except one. Also, a chess match between world champion Garry Kasparov and IBM Deep Blue’s Bot in 1997 were again a Bot won. This evidence shows us that we have to work on safer AI. Because if machines become more Intelligent we will not be able to control them and this can be disastrous. Therefore AI should help humans and not try to overpower us. Elon Musk's OpenAI is working on making such safer AI. 

Many have said that AI is the future and people who don't know about it are gonna be left behind. So, I thought to just brief about this topic. I will try to cover some of the important things related to AI in future posts. Finally, Thank you.



No comments:

Post a Comment