How do you create “Intelligence” ?

1

Posted by Noj | Posted in Uncategorized | Posted on 05-06-2011

Excellent question. Why do you want to know ? because intelligent people change the world in amazing ways. Geniuses who figured out how the world works, ruled the world. The great things we have now come from geniuses. This is a question that has bugged me for years. Is it possible to create a software that would emulate Intelligence ? is it possible to write self evolving software that would outperform manually coded software ? Geniuses fascinate and excite me. Looking at the “full list of Nobel prize” winners, TED.com as a part of my weekly media diet, one soon realises that they’re full of very intelligent people who do Intelligent stuff. If we can solve the problem of “Intelligent Software”, so many other problems can automatically be solved. This is sort of like the “Utlimate problem”, waiting humbly for someone to battle their heads against. This is also known as “Artificial Intelligence”, but I am so surprised that its 2011, and yet millions and millions of apps are made without a single drop of “AI” algorithms. The field started in 1940’s and yet my laptop and phone are both very stupid pieces of hardware and software. There is one thing that fascinates me more than anything “software that programs itself”. I’d very much like to see this happen.

Define Genius ?

According to the Dictionary it is

genius |ˈjēnyəs|

noun ( pl. geniuses )

1 exceptional intellectual or creative power or other natural ability : she was a teacher of genius | Gardner had a real genius for tapping wealth.

2 a person who is exceptionally intelligent or creative, either generally or in some particular respect : one of the great musical geniuses of the 20th century.

3 ( pl. genii |ˈjēnēˌī|) (in some mythologies) a guardian spirit associated with a person, place, or institution.

• a person regarded as exerting a powerful influence over another for good or evil : he sees Adams as the man’s evil genius.

4 ( pl. genii ) the prevalent character or spirit of something such as a nation or age : Boucher’s paintings did not suit the austere genius of neoclassicism.


Intelligence  Genome

http://en.wikipedia.org/wiki/Strong_AI – Its something we build that exceeds normal human intelligence. So how do we go about building this damn thing ? Scary but fascinating.

I’ve been thinking about this problem for months now. Here’s what I’ve come up with.

  1. life engine – The root algorithm that maximizes survival. The program needs to be running as long as possible. It should modify/replicate itself in a way that it its always stable and running.
  2. knowledge engine – model of how the world around including itself of the machine is and how it works. Its not explicit facts but probabilities of how much of a “concept is true”. The more the concepts get validated the higher the probabilities get, and new concepts start with very low probability “hypothesis stage” and then get validated to “theory stage” and then to a “law stage” as more of the concepts get validated. It can also base its trueness on other similar organisms who have been around for long and have been promoted as “authorities” by their peers (like how mums and dads work)
  3. computation engine - Due to the deep desire of survival, the software needs to be very good at predicting what’s going to happen next and orient itself to fit the situation. It needs to try millions of possibilities and calculations in its mental model and resolve to an answer that has the highest chance from the knowledge (model) it already has.
  4. perception engine – Taking input from the various input devices such as  light, sound, pain, chemicals (taste/smell) and storing this data for validation of existing knowledge and learning new patterns.
  5. response engine – output devices to interact with the world, arms, legs, speakers, and also communicating with other similar devices who could help it survive e.g learning from them, The idea is to adapt to the situation.
  6. learning engine – reflecting upon what it thought was going to happen and what actually happened. Based on this update the knowledge map. I guess the learning engine and the knowledge map are the key points to Intelligence.
  7. Communication engine – Communicate to other peers about what they have learnt, and what you have learn’t. This is to take advantage of the “Collective Intelligence” phenomenon. More like “specialize and exchange”. Loads of individuals working together to act like a much bigger and powerful individual.

If you’ve been reading the previous blog posts, this is close to what I’ve been thinking as “neuronet” or “Human Digital Cloud“.

How do we fit all these together ? I don’t know, but when I do, I shall let you know :D

Comments (1)

Great article! I’m currently majoring in computer science and AI has been something that has fascinated me for a while now.

Your seven layered approach to AI seems like a great idea.

Keep up the good work :)

Write a comment