Evolution and Neural Systems

What is the purpose of a brain and how does that effect the evolution of design?

Survival

It is obvious that the reason any animals have neural systems is because the neural system provides a survival advantage.

How does a brain provide an advantage?

The simplest neural systems are in effect hard-wired with little or no learning. These simple systems provide reactions or reflexes to environment changes that may be harmful or beneficial.  Attraction toward food or mates, and fleeing from predators is very valuable.

As neural systems become more complex and have learning on can ask: What is being learned?
The answer is often stated as "better behaviors" and such. A better formulation would be:

Brain impart a survival advantage by predicting the future state of the environment 
and pre-reacting to the environment.

Here are several examples at varying levels of intelligence:
  • A mouse is hungry and predicts a future state that includes food, and predicts where the mouse is when it can eat.  This resolves down to actions that move the mouse to the predicted location of the food.
  • The same mouse smells food and predicts from experience that moving toward stronger smell will result in a future state that includes food.
  • A rabbit smells a fox, sees a little blur of red fur, and hears paw steps.  The rabbit is then attacked and flees and gets away (barely). Later the rabbit smells a fox. The rabbits brain 'remembers' or predicts a future state of the environment that includes being attacked, and gets adrenaline pumping, muscles warmed up with tension, and alertness so it is already in a nearly fleeing state before the fox attacks.  The cost of being wrong in prediction of the future state is fairly low, whilst the cost of not being ready for an attack likely fatal.
  • A human knows that in past years winter has always come.  So predicting the next winter will arrive in 2 months, prepares food stores, a warm cave and firewood.
  • The scientist has seen gas burn and explode, has seen the lid blown off a put by steam. After running many mental simulation of many possible physical configurations, invents the piston and a gas engine. (Many many simulation and chains -of-thought)
  • Sir Issac Newton sees an apple fall and after many chains-of-thought (years!), predicts that Jupiter will be there instead of over there in the sky.
In effect brains are the great predictor, and intelligence is how deeply predictions can be made.  Interestingly chains of predictions can be a train-of-thought doing in effect infinite prediction of possible or impossible future states.

Evolution

Evolution of neural systems has proceeded by selecting the best survivors over time and creating a better and better predictor of the future, and retraining from scratch with each generation.

In a computer system the rules are a bit different.  The system can evolve without having to create a whole new system.  In an SRS one can have a mature system and then add one or more Concept Maps and a little more cortex neurons.  The subsumptive parts can be any concept one can imagine.  For example a system might have a single camera input and work well with it.  Then one can add another camera, and all the associated Concept Maps, and add in stereo Concept Maps, and continue running the system.  The new cortex and old cortex will gradually integrate in the new Concept Maps.

In effect we use neuroplasticity to let us evolve forward without having to retrain the entire system.

The system itself could use evolutionary algorithms to automatically try variants of random Concept Maps to expand the system and raise intelligence.

Comments

Post a Comment

Popular posts from this blog

UE4 Asset Directory Listing Blueprint

Image
So I needed to be able to, in game, get listing of asset directories and assets in a folder style.  The purpose is to let me make on-the-fly menu selections of something like 10000 assets in game and do it fast. Turns out in UE4.25 you can do all of this in blueprints. The low level problem is that assets get packaged into .pak files when you ship the game.  Regular C++ style file and directory operations will not find any assets.  Instead there is the Asset Manager. Assets are bundled as assets and are not in a directory structure in the pak file.  The path to an asset is in effect just a string.  So the over all flow is... Load all the assets under one or more asset directories.  When the game is packaged what was something like Content/DefaultObjects/Trees becomes /Game/DefaultObjects/Trees. When you want to ask "What sub directories and assets are in /Game/DefaultObject/Trees you call my blueprint node.  It returns a result structure with a list of string for subdir
Read more

Subsumption and Neural Hiding - The 'Object Oriented Paradigm' of Artificial Intelligence

In computer programming Object Oriented Programming (OOP) gave programmers a neat way to compartmentalize programs, rather than having a mess of disorganized code and data.  One can neatly make an object, treat it as a black-box and reduce the cognitive load on the programmer.  Nifty. In reactive neural networks where one is taking world inputs and generating response outputs Subsumption vastly simplifies the work of designing the neural processing. From dictionary.com: verb (used with object),   sub·sumed,   sub·sum·ing. to  consider  or  include  (an  idea,   term,   proposition,   etc.)  as  part  of a  more   comprehensive   one. to  bring  (a  case,   instance,   etc.)   under  a  rule. to  take  up  into  a  more   inclusive   classification. Subsumption in the context of neural networks is where some reflex is designed as a response to inputs or other neurons and overrides 'lower level' reflexes.  Then some other reflex is desired so that can then be des
Read more

Self Cycling Asynchronous Neural Systems

Current neural network systems have a general paradigm of being given some input, and then producing a resulting output. A true general AI should act like a biological system and be self-sequencing. The neural system should just free-run, responding to inputs and creating outputs and learning. One method I am experimenting with in Cognate and NuTank is having a hard-wired set of concept maps (a N-dimensional set of neurons used to embody a concept, e.g. foveolar edge angles.) that take inputs from the external world and produce hard-wired reactions as outputs to control a real or simulated robot. Such a system also uses subsumption to aid in much easier design. This is the Concept Map System (CMS) for the purposes of this post. For now lets ignore the idea that the hard-wired concept map system might also be adapting and evolving over time simulating neroplasticity. One then has a large 3D array of neurons that are in acting as a pattern matcher on the entire CMS.  What this 3D a
Read more