Intelligence requires Knowledge

“Every knowledgeable person is not intelligent but every intelligent person is knowledgeable”.

This statement provides a very strong foundation for the above mentioned Title. But first of all, what is knowledge? It is the internal representation of facts. A single fact can vary from person to person’s perspective in terms of knowledge. We will come to its detailed explanation later.

Knowledgeable person may have huge knowledge base but fails to use it when required in real life situation. But an intelligent person knows when and how to use that knowledge base to solve a problem. This intelligence makes human, different and superior from every other thing.

Artificial Intelligence also requires knowledge. But this knowledge can’t be anything. There are some knowledge characteristics:
Desirable property — Indispensability.
Less desirable Property -
1. It is voluminous.
2. It is hard to characterize accurately.
3. It is constantly changing.
4. There is a difference between the way it is organized and it is used.

AI technique is a method that exploits knowledge .
Knowledge should be represented so that -
1. Generalisation. Situation that share same property are grouped together.
2. Data provider must understand it.
3. Can be modified to rectify errors.
4. Can be used, if partially or almost complete but not entirely.

It is not yet over, there is more to it!!

Turing Test for a Thinking Machine

Turing Test is a way to test whether the machine is actually capable of thinking by itself or not. The Turing Test involves the factors like :

1. Knowledge Representation
2. Automated Reasoning
3. Natural Language Processing
4. Machine Learning
5. Computer Vision
6. Robotics

Factors from 1 to 4 are the basic requirements for a machine to be thinking. Hence, it is also known as Partial Turing Test.

Why AI needed?

In earlier days, AI was only used for game playing like Chess, Chinese Checker, Tic Tac Toe, Tower of Hanoi, etc. There was a reason behind this. To code games like Chess or other board games, all possible states had to be calculated & stored in memory if it was done in a deterministic manner. Now suppose a chess match takes 50 moves & 50 countermoves and there are 32 chess pieces(including King, Queen, Knight, Pawns, etc.). Now to calculate & store all the possibilities of this match will require 32¹+32²+……….32⁹⁹+32¹⁰⁰ bits of memory which almost Zetabytes of memory for one single match. So it was never possible to code all the possibilities of chess in deterministic fashion. That’s why AI was needed to deal with non-deterministic ways to play chess. As we have already mentioned in our previous story that AI deals with inexactness, incompleteness, partial knowledge. As this story might bore you if I provide just theoretical concept. So I have included a video of one small Windows Store App(more specifically a game, although I didn’t publish it Windows Store) made by me just to give you a hint on how much it becomes difficult to code this Tic Tac Toe in a deterministic fashion.