Skip to main content

What is Intelligence: Mind Space


The physical human brain is just hardware for simulating what I call a mind space. If we borrow an analogy from computer architecture, we will see that the human brain is just basic hardware for implementing what we can call the machine architecture, which is built from the neurons and connections, and also with all other stuff going on like neurotransmitters etc.

This brain hardware just implements very primitive stuff for implementing the mindspace. It is in the mindspace that real magic happens. On a typical physical computer, the hardware is very primitive and investigating the hardware alone without the knowledge of the software that runs on it will reveal nothing of what the true capabilities of the hardware are.

But the hardware is powerful in the sense that with very simple operations, it is able to create an environment where something like an operating system can be written. An operating system not only provides a way to coordinate the hardware but it actually provides a space for the development of applications for doing all kinds of things humans want to achieve.

This is the true power of the physical computer, the ability to provide an environment where generic computation can take place. And we must remember that computation is merely a way to represent certain kinds of problems we want to deal with. If we want to be able to simulate the weather and thus obtain predictions about when it will rain, snow, etc. the first step is to actually analyse the weather in a scientific way.

We take weather data over a period of time and analyse this data. After the analysis stage, we might have seen some underlying patterns in the weather that could be a good seed for making meaningful predictions about the weather. But then we realize that the equations are too complicated to be done by hand we resort to automation for relief. We can with some severe effort do the calculations by hand, but that will result in drudgery.

The computer comes to the rescue because that drudgery is what it was designed to automate away. So we go ahead to start trying to express our problems in a form that can be represented on a computer. In theory, we can represent our problem as a computer program on paper but to actually run it we need a physical implementation of a computer.

But like in earlier chapters where I have talked about transferring a problem into a form of representation that can be manipulated by a physical computer I have always ignored the necessary fact that the computer has its own control and coordination issues.

The computer has its own software for coordinating the activity of its parts, we cannot just take our own problem and represent it on a computer. Earlier on I took as a given that once we had a computer we could do that but in reality, the computer has a layer of software already existing called an operating system.

It is not that we cannot just use electromechanical switches to input our programs and then turn certain other switches and then read out results on the other end, in essence, using our muscles as some kind of operating system, but in modern computer the operating system is an abstraction that enables us to use simpler control logic for controlling a computer rather than building hardware to handle all levels of control.

Apart from presenting us with a mechanism for making sure that the computer is acting in a coordinate fashion the operating system also exposes for us a space where we can implement all the user based programs we want to do that are not related to the issues of control on the computer.

As for the weather program, we are contemplating, we will go ahead to write a program and run it on a real computer and then we will be able to produce results. These results can be produced as simple numbers or we can use the same computer to produce a visual representation of the simulation much like the way we could view the earth from above to see all the cloud formation, plus produce all the predictions we want in a much more pictorial fashion.

This ability to simulate the real world on a computer is a very powerful analogy with the way the human mind/brain system works. The images we see in our heads are running on the very basic hardware of our brains, and this is was what misled early AI researches. They tried to directly simulate mind activity on a computer using programming the same way we would take a video of cloud formation from above the sky and replicate it on a computer as a weather simulation without having access to some kind of mathematical model of cloud formation and the weather in general. They were concerned about the details of the mind's expression of intelligence and not the true underlying hardware algorithms that made that mind possible.

But thankfully modern AI is making inroads into these hardware algorithms with deep-learning and it is my goal in this book to inspire and also advice. We have in our hands a very powerful framework for expressing cognitive algorithms but we should look deeper for we have just come to the shores and a whole ocean of ideas await us.

The beautiful analogy of the computer and the brain is that in a computer circuit, electricity is running around and transforming one set of pattern integrities into another and the same thing is going on inside the brain too. Electricity is running from neuron to neuron transforming pattern integrities. If you look at an exposed brain or an exposed computer circuit nothing external gives off the fact of their operation. It usually requires a special tool like spectrum analysers to see what is going on inside a computer circuit and with the brain, we can take MRI images or CAT scans and see what is going on inside.

One might ask that where is the screen of the brain? the same way we look at a computer circuit and see nothing interesting with our unaided eyes, yet if we peer into the computer monitor we see the contents of the computer hardware displayed on the screen and we can control the hardware by performing actions in this screen using some input device, which is actually sending control signals to the computer hardware but we can observe the effects on the screen.

The mind is the monitor of the brain and what we see in our minds/eye, ear, tongue, etc. is actually a high abstraction of very primitive activities going on in our neural circuitry.

The brain is not a Von-Neumann computer but they are both equivalent at some level because they transform information from one structure to another which is what computation is at its most primitive form of description. The human brain and a digital computer are equivalent structures.

The Mindspace is the operating system of the brain, allowing us to run our mental programs. Not all mental programs are visualizable because being able to visualize is not necessary in many cases. Those mental programs that we cannot visualize are part of our subconscious machinery which is like 96% of all the mental activity that is going on. Our thought process, which I am generalizing as visualization is actually just a very tiny aspect of the entire mindspace/cognitive architecture.

When I say visualizing I do not mean just pictures seen in the mind, I use it as a generalization of that which we call thought, the stuff going on inside our heads. Just like the computer monitor enables us only to peer into that which is most important for us to know about in our computer systems with a lot going on that we cannot visualize because they are basic processes, our thoughts running in our mindspace operating systems only perceive a tiny fraction of what is running in our mindspace.

Thoughts are a way to interface with our internal machinery, to send queries to our database of representation, view results and create new representations consciously with imagination.

Thought is so high in the abstraction hierarchy of our minds that any attempts to try to copy human thinking in the hope of creating intelligent machines will be akin to copying screenshots of Microsoft word in operation in an attempt to create a working word processor.


Popular posts from this blog

Next Steps Towards Strong Artificial Intelligence

What is Intelligence? Pathways to Synthetic Intelligence If you follow current AI Research then it will be apparent to you that AI research, the deep learning type has stalled! This does not mean that new areas of application for existing techniques are not appearing but that the fundamentals have been solved and things have become pretty standardized.

How to become an AI researcher

Artificial Intelligence is all the rage these days. Everyone is getting on the bandwagon but there seems to be a shortage of AI researchers everywhere these days. Although many people are talking about doing AI not many people are actually doing AI research.