CODIL, Complexity, evolution and Intelligence

I have just been following a FutureLearn course

Decision Making in a Complex and Uncertain World

It was run by the University of Groningen under Professor Lex Hoogduin,. While I have been concerned with complex systems all my life I have never done even an introductory course on the subject and it has proved very useful in stimulating ideas about how I might write up my work on the evolution of human intelligence.  I felt that the information presented by Pier van den Berg on natural evolutionary dynamics and that presented by Franjo Weissing on social systems helpful as while much of what they presented was known to me their presentations help me to clarify my ideas. s a result I have posted the following closing comment (limited to 1200 characters) on the course




 

This course is proving a great help in research into the Evolution of Human Intelligence.

In 1967 research started on an unconventional “computer” with a user-friendly symbolic assembly language (CODIL). The aim was that humans and the system could work as partners on complex but mathematically unsophisticated tasks. Extensive research was done and a small package was trial marketed and got very favourable reviews. It was abandoned because of incompatibility with conventional computing technology. In retrospect a key problem was that the underlying theory had not been adequately explored.

In theoretical terms conventional computers process numbers in deterministic array of numbers while in CODIL concepts (ideas named by the human) are activated in a highly recursive network. While the original CODIL system was designed to process complex clerical-type information the recursion in the theoretical model suggest an evolutionary pathway from simple decisions at the neuron level up to the exchange of cultural information in human society. The CODIL research showed how the human brain could tackle complex tasks.

Interested to know more – see my blog www.trapped-by-the-box.blogspot.co.uk

I will be actively following up the ideas this course has generated, with various leads to follow up, and an enhanced enthusiasm for properly writing up my own research.

My personal battle between complex and complicated systems

I recently decided to drop into a FutureLearn course "Decision Making in a Complex and Uncertain World" by the University of Groningen. The opening section really made me sit up as I realized that I had never seriously thought about a formal definition that clearly distinguished between complex and uncertain systems and complicated  but predictable ones. Of course I was well aware of the difference in practice but having a definition clarified a number of issues relating to how my research into a human-friendly computer (CODIL) started, why the research came to be abandoned, and why there is now renewed interest in the subject.

Fossil Elephant Tooth

The Futile Search for the Philopopher's Stone of Intelligence

I was delighted to discover the above video about a tiny fraction of the brain of a mouse thank to P.Z. Myers. He discusses the paper Saturated Reconstruction of a Volume of Neocortex. Cell 162(3):648-61. doi: 10.1016/j.cell.2015.06.054. and points out the futility of the approach to examining the brain in ultraminute detail if the hope of understanding the basic principles by which it works.

Interestingly the authors of the paper are having doubts about the approach and write:

Finally, given the many challenges we encountered and those that remain in doing saturated connectomics, we think it is fair to question whether the results justify the effort expended. What after all have we gained from all this high density reconstruction of such a small volume? In our view, aside from the realization that connectivity is not going to be easy to explain by looking at overlap of axons and dendrites (a central premise of the Human Brain Project), we think that this ‘‘omics’’ effort lays bare the magnitude of the problem confronting neuroscientists who seek to understand the brain. Although technologies, such as the ones described in this paper, seek to provide a more complete description of the complexity of a system, they do not necessarily make understanding the system any easier. Rather, this work challenges the notion that the only thing that stands in the way of fundamental mechanistic insights is lack of data. The numbers of different neurons interacting within each miniscule portion of the cortex is greater than the total number of different neurons in many behaving animals. Some may therefore read this work as a cautionary tale that the task is impossible. Our view is more sanguine; in the nascent field of connectomics there is no reason to stop doing it until the results are boring.

My own approach, which I am developing on this blog, is to start with the idea that the problem is so complex that it is best to assume that it is infinitely complex, and any attempt to discover all the possibilities is theoretically impossible. I follow the approach used by physicists who use an "ideal gas" model because there are far too many molecules to consider individually. Instead of an infinite number of identical gas molecules with a range of kinetic energies I consider an infinite number of identical neurons. Every neuron has the potential to link with every other neuron (just as any pair of molecules can collide in the ideal gas model) and these links vary in strength. The links act as a store for the patterns of information stored in the brain, and brain activity involves passing of electrical activity between neurons - and this activity may alter the strength of the links involved. Because the model is working in an infinite framework there is no limit to the maximum complexity of memories which can be stored, and because the strength of the links change with use no two brains can ever be expected to be identical - and each brain will dynamically change with time.

The strength of the "ideal gas" model is that, while it is not perfect, it provides a predictive framework by which the behaviour of real gasses can be judged. I would be the first to admit there are limitations to my "ideal brain" model but I believe its predictions about how brains might work, and how human brains evolved, could provide a framework for understanding how real brains actually behave. This would seem a far more effective approach than some of the very expensive research projects currently underway.

Looking for the "Neural Code"

Going through an older section of my email inbox I found a Scientific American link to John Hogan's blog post Do Big New Brain Projects make sense when we don't even know the “Neural Code” in which he wrote:

Neuroscientists have faith that the brain operates according to a “neural code,” rules or algorithms that transform physiological neural processes into perceptions, memories, emotions, decisions and other components of cognition. So far, however, the neural code remains elusive, to put it mildly

The neural code is often likened to the machine code that underpins the operating system of a digital computer. According to this analogy, neurons serve as switches, or transistors, absorbing and emitting electrochemical pulses, called action potentials or “spikes,” which resemble the basic units of information in digital computers.

I prepared the following, perhaps too lengthy, comment but when I came to post it I got a message that the page had been moved - and all attempts to find it resulted in irrelevant pages on the Scientific American web site. So I am posting my response below:

------------------------------

I would argue that the problem with modern brain research is an inability to see the wood for the trees. Of course if you look in detail at the brain things get very complex. But such complexity is common in science. The key idea underlying evolution is very simple – but when you look at individual cases in detail there can be enormous complexity. The same applied in medieval times when the movements of the known heavenly bodies appeared to be very complex – until it was realised that things became much simpler if you calculated the motions of the planets using the sun, rather than the earth, as a key reference point.

The problem with the human brain is that, as John Hogan says, we don't know the “Neural Code” and virtually everyone is looking at the problem in ever greater detail – apparently on the assumption that the harder you look at the fine detail the more certain you are to find out the shape of the wood!

I have been trying to stand back and get an overview and have come up with an “ideal brain” model which in some ways parallels the “ideal gas” model in physics. All neurons (like all gas molecules) are identical – and the dynamic links between neurons are like the dynamic collisions between gas particles. Using such a simple model it is possible to “grow” a brain that can remember and use more and more complex concepts – with the complexity of the most advanced concepts it can handle depending of the brain's capacity and time for learning. The model explains consciousness and can predict detailed observations about the brain - for instance the so-called “mirror cells” in the brain turn out to be nothing special as the observations simply reflect the way that all neurons work in the “ideal brain”. In addition it is possible to ask how human “intelligence” might evolve and this approach predicts a major tipping point (rather than some major genetic “improvement”) which produces and “explosion” of “new ideas” when “cultural intelligence” becomes a more effective tool than the innate biological “intelligence” of the “ideal brain” model.

The problem with the model, and possibly the reason why it appears not to have been explored before is that to a “culturally matured” mind (and all people accessing this text on the internet will be culturally mature) the model involves several counter intuitive steps.

1. The model assumes that at the genetic level the only significant difference in the processing mechanisms between our brains and most animals relates to supercharging effects (more capacity, more links, more effective blood supply, etc.), and that if there is a difference the model actually suggests a reason why we might be genetically less intelligent that some other animals! Before you shout me down over this “outrageous claim” I should point out that the model suggests why culturally supported intelligence is infinitely more effective than the genetic intelligence foundation on its own.
2. You have to forget everything you have learnt about computers and algorithms. The definition of a stored program computer model requires there to be a pre-defined model of the task to be performed. The “ideal brain” model starts by knowing nothing about anything and has no idea what kinds of tasks it will be required to carry out. Virtually all it does is store and compare patterns without having any idea what those patterns represent. Once you start looking in great detail at how specific name tasks are processed you have taken your eye off the ball - as you are asking about what the brain can learn to do - and not what the underlying task independent mechanism is.
3. Everyone knows there can't be a simple model of a Neural Code – because with so many people are looking someone would have found it if it existed - so there is no point in looking ...
4. My research has “reject” stamped in all the standard “Winner of the Science Rat Race” boxes. I make no secret that I am 75, am not currently associated with any established research group, and the only facilities I have are a P.C. in a back bedroom, access to the internet, and access to some old research notes on a long abandoned blue sky project which was trying to design a human friendly white box computer to replace the standard human hostile black box computer everyone takes for granted.

If you are interested I hope to have a detailed description of the “ideal brain” model on my blog later this month.