If not be wrong to say that the future years are of robots. Coming to a quote that says “Within 20 years machine will be capable of doing anything a man can do”. One will imagine that this quote was stated in the year 2010 or even during the ’90s but it’s not. Actually, this was said by a computer scientist in 1960.
You all have only had about Artificial Intelligence before. Many a time it is just called AI. Whenever we come across this word we always come up with intelligent robots that are now taking over the world.
in 1954 the term artificial intelligence was coined during the world’s first AI conference. And now it’s too common to see robots working in restaurants hospitals shops and everywhere else.
NVIDIA: Learning process of robots
Despite many predictions and funding, we are still not in the position to get machines with human-like intelligence. We really need to dig into how much machines learn to see why the idea gets flop every time.
- In actuality, AI is a term that is a mix of complexity and hype. But, the central idea of artificial intelligence is that the machine could learn think and act just like humans do. Most importantly, it learns all by itself without human intervention.
- It’s so obvious that robots can not learn naturally. To overcome this challenge, scientists created neural networks late in the 1950s. One could get perplexed about the idea of neural networks. In simple and short words, these are computer programs and they make how the human brain works.
If we come to an example, say you are creating a machine or a robot to identify dogs. The very first thing which you will have to do is, bring tons of pictures including dogs into the neural network. And then the analyzing process begins. It would analyze and learn about how the dog looks. Then you can bring a real dog that is never seen by the machine it would surely recognize it.
Scientists believe that neural networks would work and they were right in the money. The problem was the raw materials needed to fuel their ambitions.
Machines learn only through analyzing data or examples. The more we provide it with an enormous amount of data, the more it is recognizable. Moving to the time of 60s and 70s, neither we had much data nor the internet was invented. Lack of data wasn’t only the hurdle. Neural networks also needed computers that could function hyperfast.
Supercomputer in 1995 was shockingly slow. This would take 7 weeks to produce 78 minutes if it worked nonstop. Imagine the situation!
Received a great Match
Two huge breakthroughs are fuelling an AI renaissance. Obviously, the first one is the internet and the other one was the revolution of Chips.
Internet handed over an unlimited amount of data. According to a recent IBM paper, it was found that 90% of the world’s data is created in just the last 2 years. Isn’t it surprising? Yes, it is. From billions of ebooks to millions of photos shared on Facebook, countless online articles, and images we now have a collection of neural networks.
The breathtaking jump is in another part of fuelling. No, it is well known that how computer chips are recognized as the brains of electronic gadgets like laptops or phones. These chips are complex and contain transistors known as brain cells. Do you know, there are nearly 8.7 billion transistors on the chip of the latest iPhone. Similarly 117 Sun Microsystems contain 1 billion transistors.
Remember the blocking graphics on video games like Mario and Sonic?
- If you look at those graphics now you will find that they have become more realistic since then.
- All the credits, to this incredible jump go-to chips which are called graphic processing units i.e GPU.
- it is capable of performing many calculations at one time which helps in creating movie-like graphics. And here, is the difference between how traditional chips worked which carried calculation one by one.
NVIDIA: Most important company in America
Artificial intelligence or AI has become a buzzword in this tech world. According to data from Bloomberg, it shows that about 840 US firms mentioned AI in their recent earning reports at least once. Coming to the reality it says that few of these companies are really building intelligent systems.
With the booming AI business, there’s only one company NVIDIA (NVDA). Back in the1990s, it invented graphics processing units and is responsible for realistic video game graphics that we come across nowadays. And then an important Discovery was made. These gaming chips where are absolutely perfect for the neural networks of robots or artificial intelligence.
Very soon, NVIDIA started making chips specifically for machine learning. it’s really amazing to know that the AI-related sales in the first half of 2020 which was at the top with $2.8 billion. If you come to the present scenario we will find that 90% of the neural networks are running on NVIDIA GPU (graphics processing units).