p-nazariy / lu_brain

Lu_Brain: the new awesome way to AI

Home Page:https://www.ludyna.com/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Lu_Brain

Lu_Brain is an artificial neural network that is not a gradient descent algorithm. But, of course, human brains do not use gradient descent algorithms, too; that would be a big waste of computational and energy resources.

More info soon. This project is a work in progress at this moment.

About

I believe human brains are optimal or close to optimal. So to create natural intelligence, we have to create something similar (in an algorithmic sense) to what our brains do.

I see two main approaches to achieve it:

  1. Try to understand brain algorithms by observing them. Unfortunately, this approach is tough to deduct complicated algorithms. It's like trying to understand what computer programs do by only observing hardware (CPU, motherboard, RAM, etc.). This approach is good at understanding basics, but it is hard to deduct algorithms from it. Real neural networks are much more complicated than modern CPUs.

  2. Try to use logic, imagination, creativity, intuition, math, and what we already know about our brains to recreate our brain algorithms.

In this project, I use the second approach while taking basics from the first one. To invent algorithms in this library, I asked myself how nature would do this, how the brain could evolve to what we have now, what baseline principles are, etc.

What is different about this project compared to purely mathematical models based on gradient descent algorithms is that the resulting library contains an algorithm(s) close to what the human brain does, in my opinion. Thus, after I figured out baseline principles, there was only one right way forward. However, I think only knowing how part of the brain work is not enough. So I spent a lot of time translating that knowledge to an algorithm that can optimally work on our computer systems which are very different from a highly distributed brain.

Why

  1. To better understand how human brain algorithms work, which has application in medicine and health-oriented devices and making humans more intelligent.
  2. To open possibilities for new tech like neural databases, artificial telepathy, clever robotic, and virtual servants.
  3. Move AI's ease of use to the next level.
  4. Bring tech that will help humanity to survive the next extinction event.

FOR THE GLORY OF HUMANITY AND THE LARGER CONSCIOUSNESS SYSTEM! :)

About

Lu_Brain: the new awesome way to AI

https://www.ludyna.com/

License:MIT License


Languages

Language:C 100.0%Language:Shell 0.0%