How to solve the problem of Artificial General Intelligence (AGI). Part - I : Introduction

An algorithm is always a heuristic to solve a computational problem.Reduction of search space to speed-up calculation is of primary importance in computer science. Albert Einstein said that constructing a problem is much more important than finding it's solution which is merely a technical matter.The main problem of artificial General Intelligence (AGI) is to find a universal algorithm to construct correct problems.

We need to find out why Artificial Narrow Intelligence (ANI) is comparatively easy but AGI is difficult.What prevents one specialized algorithm to solve another domain's problem? What is the fundamental difference between problems of different classes?We know that diagonalization procedure can reduce any NP-Complete problem to another. Is it possible to use that for creation of useful algorithms?

There may be a phase transition between ANI and AGI related problem classes with different order parameters. We could dynamically create algorithms to solve problems if the structure of the problem assists us to create the algorithm so that desired solution could be achieved.

Intelligent behavior could be described as a class of problems where correct behavior is difficult to produce but easy to verify. The solution may be found in the way we understand this issue because according to Werner Heisenberg we need to define first what we mean by the term 'understanding'.