On Intelligence [Paperback]

We often routinely talk about intelligence and we attempt to measure it for for a variety of purposes. But do we know what it is? Jeff Hawkins is one of the first people to present a specific and comprehesensive theory of intelligence with a leading role for the human neocortex. Hawkins starts by stating that Human intelliigence is fundamentally different from what a computer does.

But isn't artifical intelligence (AI) a good metaphor for human intelligence? No, says Hawkins. In AI a computer is taught to solve problems beloning to a specific domain based on a large set of data and rules. In comparison to human intelligence AI systems are very limited. They are only good for the one thing they were designed for. Teaching an AI based system to perform a task like catching a ball is hard because it would require vast amounts of data and complicated algorithms to capture the complex features of the environment. A human would have little difficulty in solving such everyday problems much easier and quicker.

Ok, but aren't neural networks then a good approximation of human intelligence? Although they are indeed an improvement to AI and have made possible some very practical tools they are still very different to human intelligence. Not only are human brains structurally much more complicated, there are clear functional differences too. For instance, in a neural network information flows only one direction while in the human brain there is a constant flow of information in two directions.

Well, isn't the brain then like a parallel computer in which billions of cells are concurrently computing? Is parallel computing what makes human so fast in solving complex problems like catching a ball? No, says the author. He explains that a human being can perform significant tasks within much less time than a second. Neurons are so slow that in that fraction of a second they can only traverse a chain of 100 neurons long. Computers can do nothing useful in so few steps. How can a human accomplish it?

All right, human intelligence is different from what our computers do. What then is it? I'll try to summarize Hawkin's theory.

The neocortex constantly receives sequences of patterns of information, which it stores by creating so-called invariant representations (memories independent of details). These representations allow you to handle variations in the world automatically. For instance, you can still recognize your friends face although she is wearing a new hairstyle.

All memories are stored in the synaptic connections between neurons. Although there is a vast amount of information stored in the neocortex only a few things are atively remembered at one time. This is so because a system, called `autoassociative memory' takes care that only the particular part of the memory is activated which is relevant to the current situation (the patterns that are currently flowing in the brain). On the basis of these activated memory patterns predictions are made -without us being aware of it- about what will happen next. The incoming patterns are compared to and combined with the patterns provided by memory result in your perception of a situation. So, what you perceive is not only based on what your eyes, ears, etc tell you. In fact, theses senses give you fuzzy and partial information. Only when combined with the activated patterns from your memory, you get a consistent perception.

The hierarchical structure of the neocortex plays an important role in perception and learning. Low regions in the structure of the neocortex make low-level predictions (about concreet information like color, time, tone, etc) about what they expect to encounter next, while higher-level regions make higher-level predictions (about more abstract things. Understanding something means that the neocortex' prediction fits with the new sensory input. Whenever neocortex patterns and sensory patterns conflict, there is confusion and your attention is drawn to this error. The error is then sent up to higher neocortex regions to check if the situation can be understood on a higher level. In other words: are there patterns to be found somewhere else in the neocortex, which do fit to the current sensory input?

Learning roughly takes place as follows. During repetitive learning memories of the world first form in higher regions of the cortex but as your learn they are reformed in lower parts of the cortical hierarchy. So, well-learned patterns are represented low in the cortex while new information is sent to higher parts. Slowly but surely the neocortex builds in itself a representation of the world it encounters. Hawkins: "The real world's nested structure is mirrored by the nested structure of your cortex."

This model explains well the efficiency and great speed of the human brain while dealing with complex tasks of a familiar kind. The downside is that we are not seeing and hearing precisely what is happening. When someone is talking we by definition don't fully listen to what he says. Instead, we constantly predict what he will say next and as long as there seems to be a fit between prediction and incoming sensory information our attention remains rather low. Only when he will say something, which is actively conflicting with our prediction, we will pay attention.

The author takes his model one step further by saying that even the motor system is prediction driven. In other words, the human neocortex directs behavior to satisfy its predictions. Hawkins says that doing something is literally the start of how we do it. Remembering, predicting, perceiving and doing are all very intertwined.

I think this is a fascinating and stimulating book. Many questions about intelligence may remain unanswered but I believe this book to be a step forward in our quest to understand intelligence. The author predicts we can soon build intelligence in computersystems by using the principles of the neocortex. He is optimistic about what will happen once we succeed in this. He (reasonably convincing) argues these systems will be useful for humanity and not a threat.

Coert Visser, www.m-cc.nl

This book explores the question "what is intelligence?", introducing rather surprising results of brain research in a way that should be understandable to a person of average intelligence.

Mr. Hawkins explains that intelligence is not what people usually think it is. For most of us (including me prior to reading this book), the word "intelligence" associates with "artificial intelligence" and that associates with computers that beat humans at chess, and begs the question: will computers smarter than us take over the Earth one day?
Of course, I have always noticed, as probably all serious chess players have, that computers play chess in a totally different way. In a nutshell, they tend to excel at short-term tactics while being ridiculously inept at endgame in which one needs a good grasp of long-term strategy. Humans often make the kinds of mistakes chess software would almost never make, and chess software often makes mistakes a human player would hardly ever make. I have, though, never been able to figure out what it is that makes the "artificial intelligence" so different from "human intelligence".
That's where this book steps in. It's not about chess, of course - I just brought that example to help you understand what the problem's all about. This book is about scientific research that convincingly refutes the popular belief that computer CPU's are "electronic brains", or that a human brain is just a very, very compact, very, very efficient computer.

I shall now try to explain briefly how this book describes the essential difference between brains and computers.
In an electronic circuitry, you have a certain module (or a network of modules) that turns a certain kind of input into a certain kind of output. For example, every time you press "escape" on your keyboard, a certain signal goes into the CPU that decides what to do about it. Every time there is a certain input, there will be the corresponding output.
Brain research reveals that a brain works in a different way. It isn't like each time you walk into your room, you think "there's the window, there's the desk, there's the chair etc". Your eyes would see all those things but you wouldn't be paying attention to them because they are as usual. But if someone removed the chair, you would enter the room and instantly notice THAT THE CHAIR ISN'T THERE. It's not like you think "there's the window, there's the desk etc", and you would have to somehow deduct that the chair isn't there. On the contrary, the chair's NOT BEING THERE would be the first thing you'd notice.
So it would appear that certain parts of the brain monitor certain aspects of the outside world, getting data from their incoming neurons, and fire their outgoing neurons ONLY when something in their input data was NOT AS EXPECTED. (In the book, you'll get a detailed description how this physically works with brain cells with various functions organised in various parts of the brain and various layers of brain tissue.) In other words, intelligence isn't calculating, it's predicting. Based on past experiences, the brain sets up expectations, and reacts when sensory input isn't as expected.

You'll learn other amazing things about the way the brain works differently from what most of us have imagined so far, but I don't want to go into too much detail here.

One thing in particular that fascinated me about this book was the way Mr. Hawkins illustrates highly abstract and technical concepts with simple examples. He says he does so because he likes simple things. So do I. Two thumbs up for that.

Mr. Hawkins's message is: the computers the way they are designed today can never become intelligent because the brain works in a totally different way. Mr. Hawkins insists that really intelligent machines can only be built by imitating the structure of the brain. (I would like to ask him if he thinks that such machines would obey our orders, but that's not the point right now.)

This is one hell of a remarkable book. In general, I am not too fond of neither philosophy nor abstract science. I prefer books that contain something that I can actually use to improve something in my life. That said, the abstract science in this book is so fascinating that I consider it one of the most valuable books I have ever read, even though I won't be able to put any of it to immediate practical use.

"Prediction is not just one of the things your brain does. It is the primary function of the neocortex, and the foundation of intelligence." (p. 89)

Perhaps the crux of Hawkins's insight into how our brains work and how that is different from how computers work can be gleaned from considering how to catch a ball in flight.

It used to be thought that such tasks were solved by the brain through calculation. The brain would calculate the flight of the ball, adjusting the muscles of the body appropriately so as to arrive at a spot where the ball would be and grab it. Artificial intelligence people working on robots used this method and found out that it was enormously complex, so much so that the robots remained clumsy (and not about to play centerfield for the New York Yankees).

What Hawkins is saying is that the brain does NOT calculate the flight of the ball but instead recalls from memory similar flights of balls while at the same time recalling again from memory the muscular workings of the body as it went after and caught or did not catch similar balls in flight. After a bit of practice (storing memories) a person can get very good at catching balls.

In other words the brain predicts where the ball is going to be not through a laborious and lengthy calculation but through memories of similar events. This is a startling insight. Hawkins shows how everything we do is based on our brain's ability to predict events based on previous experience. Here's how it works:

First there is a "sequence of patterns" of past events stored in the brain.

Second, the brain has an "auto-associative mechanism" that allows it to "recall complete patterns when given only partial or distorted inputs." (p. 73) Unlike computer intelligence, human intelligence can figure out that "Wass up?" means the same thing as "What's up?" or that a face seen from one angle is the same as that face seen from another angle or even seen in some sort of distortion. This is something computers cannot reliably do.

Third, the brain stores "invariant representations" of things seen, heard, felt, etc. "Invariant" in this context means unaffected by differences in light or tone or inflection or background or any one of millions of small, inessential differences that could throw us off. These representations are not exact. They are in a way like Plato's ideal forms except they are not ideal but generalized. They are memories of the relationships between and among various features. In the case of a human face, Hawkins writes that what makes a face recognizable "are its relative dimensions, relative colors, and relative proportions, not how it appeared one instant last Tuesday at lunch." (p. 81)

Hawkins's definition of intelligence in terms of predictive ability is what I found most exciting in the book. When people talk about intelligence I usually want to demand "intelligence for what?" since the criteria for defining intelligence has always been so muddied. One of the ways of establishing a theory in science is through its ability to make accurate predictions. To judge the brain the same way seems strikingly right. Not only that but no longer do we have to beg the question of what intelligence is. It is the ability to predict.

These predictions are about everything in our lives and they involve all of our senses. As Hawkins puts it, "All regions of your neocortex are simultaneously trying to predict what their next experience will be. Visual areas make predictions about edges, shapes, objects, locations, and motions. Auditory areas make predictions about tones, direction to source, and patterns of sound. Somatosensory areas make predictions about touch, texture, contour, and temperature." (pp. 88-89)

While the first five chapters are eminently readable and exciting, Chapter 6, "How the Cortex Works" (the longest in the book) might be a bit tedious and technical for the general reader. (I know it was for me.)

In Chapter 7, "Consciousness and Creativity" Hawkins writes, "Most of what you perceive is not coming through your senses; it is generated by your internal memory model." (p. 202) We do not experience the world directly and we do not interpret it objectively. Our predictions in a sense are prejudices or stereotypes that sometimes lead us astray. Hawkins writes, "...you could substitute the word 'stereotype' for 'invariant memory'...without substantially altering the meaning. Prediction by analogy is pretty much the same as judgment by stereotype." (p. 203)

In the final chapter, "The Future of Intelligence" Hawkins makes it clear that intelligent machines will not be taking over the world. He writes, "The computer in your home, or the Internet, has as much chance of spontaneously turning sentient as does a cash register." (p. 214) Furthermore, an intelligent machine "will not have a mind that is remotely humanlike unless we imbue it with humanlike emotional systems and humanlike experiences. That would be extremely difficult and, it seems to me, quite pointless." (p. 208). Finally, fears that machines will take over the world "rest on a false analogy...a conflation of intelligence...with the emotional drives of the old brain--things like fear, paranoia, and desire. But intelligent machines will not have these faculties. They will not have personal ambition. They will not desire wealth, social recognition, or sensual gratification. They will not have appetites, addictions, or mood disorders." (p. 216)

Hawkins goes on to predict that, with an approach based on learning and memory instead of brute calculation, we will build truly intelligent machines, the applications of which will be numerous and include applications impossible to predict.

I would like to point out that Hawkins' idea that our cortex is continually making predictions about the environment, predictions that we scarcely notice unless they are wrong, is similar to an idea that John McCrone presented in his book Going Inside: A Tour Round a Single Moment of Consciousness (2001), a book I also highly recommend.