This is the second post of the series in AI vs Humans, In this post, we will discuss on why AI are smarter than and what makes them smarter than humans.
- How the human brain interprets data – AI vs Humans
- What makes an Artificial Intelligence Intelligent – AI vs Humans
- Artificial Neural Network and Evolutionary Computation – AI vs Humans
Assuming we know about what is an Artificial Intelligence, why you think they are smarter than us ? Is it because they can compute fast or since they are machines so, they don’t get tired ? To an extent yes. Let us list out the basic things why AI is smarter than us and then get into discussion of it and keep comparing it with Human brain too.
Artificial Intelligence are smarter because of :
- High Computational speed.
- Complex and efficient algorithm which forms the brain.
What is the computational speed of AI ? Is it a match with Human Brain ?
MIPS or million instructions per second is used to measure the computing speed of computers.1 MIPS can extract simple jobs, for example, tracking a white line that too on simple background. 10 MIPS can be used in smart bombs or self driven vehicle. For a three dimensional recognition, it requires around 1,000 MIPS and so on. It takes a lot for a machine to do even a simple task, thus requirement of hardware and efficient algorithm increase If we have to make a comparison we can relate the nerve volume to computation of AI.Human eye or Retina to be more specific and there are 5 types of neurons(Yeah its amazing). All the neurons which join at the end form optical nerve, which carries the signal to the brain. They react differently when they see an image and its passed to the brain to compute it. Overall, the retina seems to process about ten one-million-point images per second. Thats the reason, why we are able to pick up images unknowingly. You can say its a camera with some amazing megapixels.
So now how fast AI or a robot can do ? For a single motion detection, it needs 100 computer instruction. So 100 million instruction for a million detection of motion and repeat it 1000 times to get to the power of retinal function. Now that’s too much.
If we start thinking of make a machine like this, you can make it but at the cost of 100 million dollar project. Unless the cost, of hardware and efficiency of the machine is overcome it wont be cheap. Then there is another catch, The rapid development of computer hardware and technology holds the promise to deliver that in 20 years. Are you ready ?
Complex and efficient algorithm
There is an algorithm which makes AI to be smart. They are right from apply reasoning capabilities to reach a conclusion, systems with very strong pattern recognition capabilities, techniques for reasoning under uncertainty, applies biologically inspired concepts which can lead to evolution of robots or AI for-maintaining their own population and evolve with time.
I have selected two algorithms to explain, one will be Neural Networks and other is Evolutionary Computation. So the next post will be dedicated to these to algorithms.