Keene Macheene

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...

In a Subsumptive Regular System (SRS  here )  the  regular array of neurons (cortex) that has the goal of predicting the future state of the subsumptive (cerebrum) system . If the match is strong on the present state of the subsumptive system then the stimulus of the subsumptive system into the predicted state is strong.  In the event of sudden changes in the actual external environment one wants to ignore the predicted state and react to the immediate environment. This has an evolutionary advantage of pre-reacting to the environment and getting a jump on surviving the future state. This document calls out the various axis of selection for algorithms.  Research has yet to show where the decision points are optimal. There are several decision points on the algorithms to select: Connectivity from the Subsumptive to the Cortex The subsumptive part of the SRS is hard-wired to take inputs, compute concept mappings (abstractions) of the external environment an...

Just a simple thought: The Feathers Problem One of the challenges in the invention of powered flight was deciding what matters most.  If one wants to imitate birds there are many factors: Wing curvature Wing warping or twist Flapping Feathers Jumping into flight from the ground Energy source and many more. Some early pioneers of flight proposed feather covered wings.  After all, that is what birds have.  Flapping was tried also. In the end the Wright Brothers figured out that the salient factors were lift, drag, thrust, and weight. The lift was generated by wing curvature, not the fact there were feathers. So the first working airplane (and first gliders that worked) did not have feathers. The other solutions they came up with were light weight engine, wing warping (flaps came later) for control, no flapping but instead propellers, linear sliding take off rather than a jump, solved the learning cure issue with learning to fly a craft without any craft worki...

Subsumptive Regular System (RSR): A subsumptive spiking neural network with inputs, hard-wired processing, and outputs; combined with a regular (uniform)   3D array of learning and pattern recognizing neurons that is trying to predict the next state of the subsumptive, all asynchronously free running. The subsumptive system is the cerebrum, and the regular system is the cortex.

Got the ConceptMap editor working such that one can create Concept Maps, drag connections and make connections.  Also have the ImageInputDrive (a sub-class of Concept Map) pretty much working. Got stuck for a bit in some WinForm layout issues and scrolling the editing area.

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 ...

Neural systems in the style of Convolutional Neural Networks et. al. use matrix math and tensors to simulate the overall activity level of neurons. The reason for "overall activity level" rather than individual pulses from neurons is that the overall activity is simply the integration over time of what would be pulses in real neurons. Overall activity also is amenable to calculus with derivatives for back propagation. The reason for matrix math is locality of memory reference and simplicity of written equations for what the system is doing. Microprocessors have local CPU memory cache, and main memory. Getting data from main memory to local cache is very slow relative to processor instruction times. Matrix math tends to make operations happen in address sequential order so is efficient in cache.  Also GPU design favors sequential and matrix operation and is ultra-fast. All the above makes tensor style processing for AI efficient. Pulsed neurons are more object oriented, s...

Working on the editor.  Writing in C# so using WinForm style UI code. Got Concept Map all draggable as a UI element.  In the middle of Connections in the UI.

So, I got past the hurdle in the ConceptMap and ConnectionDef code where one N-Dimensional map can be connected to another N-Dimensional map. The connection can have source and destination ranges of map points on any and all dimensions, with weights interpolated forward or backward on all dimensions.  This ends up being a nasty confusion of interpolation, except if you make single dimensional iterators. Then you nest the iterators and it works. Next is getting a first cut UI working in Win Forms C# to edit maps and connections, and visually show activation levels and firing 'sparkle'.

A blog.  YAB (Yet Another Blog) I lost any ability to login or recover ownership of my blog Thunderfist Podium . Oh well. The intent of this blog is to flesh out some ideas on Machine Learning, Artificial Cognition, Neural Nets, Spike Neurons, and probably some philosophy thrown in for good measure. Waaaaay back in 1985 I published some peer reviewed work on Artificial Cognition. Also I made a program/experiment/game called NuTank written for DOS VGA in Turbo C. I still have the source and NuTank.exe and it even runs in DosBox . Now computers have caught up in power to actually try solving the problem. Now I am rewriting the core libraries and the whole system in C# and .Net as several packages. Cognate - The core Concept Map and Cognition/Neural Net/Pulsed Neuron Model. This has the run-time code and the ability to load/store cognitive systems, and the design level definitions.  And much more. NuTank - A (Steam platform?) C# game, multiplayer where 'tanks' battle...