The Hypothesis I Want to Test Throughout my Lifetime Research

1. Our intelligence is based on our ability to identify patterns:
   Self repeating structures within or across different domains (=abstract), in space and time.
2. There are two strategies to find them:
   1. Trial and error
   2. Use the patterns that were already found in a different domain, or on a meta-level.  (Sometimes the change of domain is confused with a change in dimension, notice that my definition of “meta” is “exponential”)
3. Patterns are usually successful models of only a segment of the observed real world.
4. Symbols and formulas are compressed ways to express the regularities that define a pattern.  They work as keys or indexes within our brains and we use them to recreate (decompress) the pattern.
5. The superposition principle is the one that allows us to express more complex patterns as compositions of simpler ones.
6. When a phenomenon is too complex to be modeled with a regular structure (pattern), we eliminate part of it and qualify it as “noise”.  By removing the noise, we leave a structure that can be dealt with, with the steps exposed in the previous points.  This approach is just an approximation and it can not be clear when it has been done in the “less bad way”.  As children, we tend to look for new patterns; as adults our tendency is to identify patterns through recognition: association with patterns that exist already within our brains (biological neural networks).  We test our models (proposed fitting patterns) according to their predictive ability at different precision levels.
7. Some initial structures and methods are inserted into us by our genetic code.  Some may be there just to provide us with some advantage.

Questions:

• What is the minimum required to start?
• How do we get rid of the noise?

Strategic questions:

• How many of these hypothesis are correct?
• How to start studying them?

GEB map problem… for explaining!!!

Gödel, Escher, Bach: An Eternal Golden Braid is a prototypical book for Artificial Intelligence.  Among the huge variety of interesting topics in it, it deals with the problem of transferring knowledge from one brain to another, through language.  Basically, the problem is that knowledge is encoded in nets built from experience and, every person has built their own nets in many distinct ways.  While trying to transfer a concept it is required to establish some sort of mapping between the net in the brain of the transmitter, to newly generated connections in the brain of the receiver on top of whatever this other brain has.  It is not always clear how difficult it will be to achieve that.  In education this problem has a terrible manifestation as students of one class have different backgrounds, since the teacher will have to adapt his/her explanation to ALL? backgrounds… No need to mention, it gets bad when one student must explain something to a group of different teachers :S  Tip: make sure ideas are as clear as clean water.