Slashdot Mirror

Check out this great book by Jeff Hawkins, creator of the Palm, called On Intelligence. His work is about understanding how the brain really works so that you can make truly intelligent machines. Fascinating stuff and firmly based in the facts of reality, which is refreshing to say the least.

I highly recommend reading Jeff Hawkins's On intelligence which proposes a structural model of the brain based on the concept that, at every level of the cortex, there is only the one operation of pattern recognition and repetition being performed. It doesn't attempt to model 'old brain' (emotions, life support, etc.,) just the cortex and intelligence itself.

The model puts forth hypothesis about a number of interesting things that one can notice about their own operation. Hawkins proposes that at every level of processing a cluster of neurons attempts to recognize the pattern and predict what will occur next. This is theoretically why one might never pay any attention at all to something like the way their front door looks but will immediately notice if there's a spot of paint on it that wasn't there before; a "pattern mismatch" signal gets passed up the chain of command from subconscious to conscious. This is why people can't help but look repeatedly at facial deformations; the face is so very familiar that any small irregularity sets off alarms. After we're used to the new pattern, it fades back into the background. After living in a dirty room for a month, it seems less cluttered and more 'the way things are'; after smelling a particular smell for a while it no longer smells like anything in particular.

An absolutely fascinating read. That book is the reason I'm going into a CE doctoral program to study that sort of thing.

I actually think Jeff Hawkins was on to a great definition of intelligence with his book that came out a few years ago, On Intelligence. Intelligence is just the ability to predict things. Every time you make a move or say something, it is in anticipation of the predicted results. Computers that spit out answers a la the Turing Test or Big Blue but can't assess the affect of what they do aren't intelligent. And that ability to predict can be answered by the simple question why--why did you do that? ("Why, Mr. Anderson!? Why!?) A computer that could answer that question, or be analyzed in such a way to find that answer should be considered intelligent.

Not mystifying at all if you subscribe to the theory that the brain is (mostly) nothing more than a pattern recording and playback system, as posited by more than a few AI researchers, including fairly recently by Jeff Hawkins, in his book "On Intelligence", wherein he describes a method by which we may be able to successfully replicate the cerebrum in hardware/software for a variety of tasks.

If our brain (and by association, our "self") is nothing more than such a mechanism (in the form of a very complex and highly connected neural network), the fact that certain neural pathways and nodes which identify certain patterns for speech can be lost, while others can still work, isn't that much of a mystery - it is just a routing issue. Perhaps the brain, on sensing irritation/inflammation connected with certain types of speech (perhaps everyday speech), routes around those speech patterns as a protection mechanism to allow healing of the vocal tract, when the user won't just shut up and quit irritating them. However, over a protracted length of time not speaking (or speaking "abnormally", such as singing), those pathways/nodal "weights" (however these connections between neurons, axons, and dendrites work - I am speaking of weights in the form of a neural network, which is simply a model, of course) get lowered to the point where the route is completely lost, causing the condition - even though the vocal tract is now healed and could be operated normally. Finding another route (in essense, learning to speak again, like baby's first words) for the neural pathways (or strengthening the old route where possible) seems perfectly possible, and obviously (in the Scott's case) doable.

What I want to know is why (seemingly?) nobody has tried this approach before (hell, I would have thought about it had I known and taken an interest in the problem - I honestly thought it was a physical issue with the vocal cords, not a neurological problem)?

Finally - this is obviously a good reason why you should not continue trying to speak when you have "lost your voice" or are hoarse - you may actually lose your voice, permanently!

Number one, would we know such an intelligence if we saw it? Furthermore, could we determine what such an intelligence was thinking if we could? I think we could deduce that something was behaving in an intelligent manner, if we applied the right tools to the purpose. The danger would be in not knowing if the object we are studying feels that such testing is against its interests and acts to stop the testing. For example, a theory of emergent behavior within large groups of people (think large bureaucracies or societal constructs) might indicate the possibility of a "group" or "hive" mind arising from interactions between the individuals involved, that is both of the individuals yet outside of it (same as Mind is to Neurons). The output of such a "mind" might seem to be intelligent, but is there any way to actually know what it is thinking, or how it is communicating? Can a neuron ever know of the human, or brain, or mind? What would we (as a human) do if a neuron suddenly could understand? Is it in any way possible such a group mind would act in the same manner? Would we understand it if it did? Can we assume that such a thing isn't happenning already in our increasingly connected and interacting world?

Secondly, I think the other problem with building such an artificial mind is that of design and construction. Interestingly, we likely have both at hand. For design, I subscribe to the view that the mind (or at the very least, the cerebrum) is nothing more than a pattern recording and playback machine, as detailed by Jeff Hawkins in his book On Intelligence. I am pretty certain that this idea is spot-on, and is something that should be investigated much further. As for construction, the design of Dr. Hugo de Garis's CAM-Brain Machine (CBM), as realized by Genobyte, seems to be the approach to use to build a system similar to what is described in On Intelligence. These machines were actually built, shipped, and used in a few research institutions around the world. Whether they still exist or not, or are buried in a back room, is anyone's guess. The fact is that they aren't a standard design for a computer, and furthermore they utilized Xilinx FPGAs that isn't manufactured anymore (whether a similar machine could be built using a different Xilinx FPGA is another matter), leads me to wonder what will happen to these machines as they end their useful lives and/or have hardware failures. Also, it doesn't appear that Genobyte is in business anymore, though their website still maintains "ghostship" status.

Maybe I am reading too much into either of these ideas? Maybe both are a bunch of hooey (indeed, the whole CAM-BRAIN machine thing is something that I am not sure whether to completely believe or not - I seem to remember a /. article a long time ago in which another company linked to this - STARLABS - was seen to be a hoax or something?). Even so, the ideas seem sound, even if the implementations don't exist in fact (although, all the research I have done seems to indicate that these systems do in fact exist).

Sounds like a cross between what Jeff Hawkins described in On Intelligence, and the FPGA evolvable hardware of the CAM-Brain Machine project...

After much reading of liturature in various fields of artificial intelligence, emergence, chaos theory, network theory, psychology, etc - I have come to believe that recognizing this is paramount to understanding how the human mind works (as well as how it doesn't). Those with Asperger's and others on the autistic spectrum can use this knowledge, along with reason and logic, to almost scary benefit to themselves, if they so choose.

In a way, proper use of this knowledge is directly akin to application of the ideas inherent in NLP (Neuro-Linguistic Programming). In essence, since you, as a non-NT individual, may need to be able to conciously "see" (and learn to understand) others emotions, as well as conciously "playback" similar emotions based on other cues - you have at hand a very useful tool to control other people, if you can understand and master it. Because those patterns you playback will cause the other individuals or group to recognize (in an unconscious manner, for most) the pattern, and play back others. What we think of as "free will" is not so much randomness in what we think we are doing, but pattern recognition and playback of patterns in a network of complex emergent social interactions in an environment of (mathematical) chaos. Non-NT individuals are among the only ones in a group who can use this knowledge to their advantage. In NLP training (which is something which seeks to train non-NT individuals to recognize what an NT individual already has practice with), the goal is recognize these body, facial, and verbal cues of others, and to utilize your own, in a manner to direct and control the responses of other people.

Done right, what can be accomplished seems amazing. With enough practice, you can almost get people to do things for you that they themselves wouldn't ordinarily do, and they do them willingly! Mind control? Jedi-like power?

Not at all - you are simply using the pattern recognition and playback capabilities within the mind of a human organism(s) in a manner concious to yourself. Hacking others brains, emotions, and desires, if you will. Some might call that unethical. Why is it ok if everyone (NTs) already do it unconsciously, but not if you (a non-NT) do it consciously? In the end, it doesn't matter, because they can't really control it (unless they are non-NT as well, and recognize what you are doing).

Your wish is their command...

Perhaps SkyNet is not far away. Or maybe Kevin Kelly should stop watching old Terminator movies.

The idea is that a computer is intelligent if it can hold a conversation with a human such that it is indistinguishable from a conversation with a real human.

RIDICULOUS.

Have you ever actually tried talking to one of these bots (including ALICE)? It is very easy to know that you're not talking to a human. Exceptionally easy. The Loebner Prize judges consistently grant the bots handicaps, acting as if they're actually being fooled. Obviously they're not, and the AI community just wants people to think that it's more advanced than it really is. Unfortnately, some members of the public *are* fooled by that.

The problem is in the Turing Test itself. It assumes that the measure of intelligence is humanoid conversational ability. I strongly disagree with that. Conversation ability is no measure of intelligence. Just for an example, I am exceptionally intelligent (statistically), but I am a poor conversationalist. Casual small-talk has always bewildered me. If I entered myself into the Loebner contest, they might think I'm a bot. Hell, ALICE might accuse me of being a bot.

Anyone who's taken an IQ test will recall that every last question has something to do with pattern recognition. You'll also recall that you were not asked to respond to any conversational questions. That's because invariant pattern recognition abilities (in a loose sense -- this also includes memory/learning and inductive reasoning) are the true mark of intelligence, and this is nearly undisputed. If they really want to test how intelligent a program is, they need to test its patern recognition ability.

Take this program -- http://www.stanford.edu/~dil/invariance/ -- for example. It's gone largely unnoticed, yet it is concrete proof of a huge breakthrough in computer intelligence. This is a little Matlab demo of a very abstract multi-layer intelligence algorithm. In this particular implementation, it is taught a set of small images. Then you can play "Pictionary" with it, drawing shapes and have it recognize them. You may say that this is unremarkable, that shape-recognition is a trivial algorithmic matter unrelated to intelligence. But the author noticed that he could draw shapes "incorrectly" -- like, the little duck picture, except with its head missing, or alphabetical symbols rotated or flipped -- and the program still recognized them. (It failed a few times, but in situations where the shape is so mangled that I would have probably failed too. How's that for a Turing test?) And this program's genius lies in not what it does, but how it does it. All of its functionality is completely abstract. It is a pattern recognizer, not a bitmap-tracer, and there are no hard-coded routines for checking if the image is flipped, rotated, etc.

This is what Palm/Handspring founder Jeff Hawkins (also the founder of new neuroscience startup Numenta, http://www.numenta.com/) calls "Real Intelligence," to distinguish it from the failed Artificial Intelligence effort. He feels that the right way to make computers intelligent is not to have them outwardly imitate human behavior, but to internally function the way the mind really works. Anyone interested should check out his book, On Intelligence http://www.onintelligence.org/. You'll wonder why you ever believed the AI hype.

Artificial Intelligence is a sham, by its very nature. Real Intelligence will be the way of the future.

As the submission noted, this work will be building on what Hawkins wrote about in his recent book, On Intelligence. The companion web site for the book is here:

There are also a some reviews of the book:
http://blogger.iftf.org/Future/000605.html
http://www.computer.org/computer/homepage/0105/ran dom/index.htm
(By Bob Colwell, who was Intel's chief IA32 architect)
http://www.techcentralstation.com/112204B.html
http://www.corante.com/brainwaves/archives/026649. html

A quote from his book:

The agenda for this book is ambitious. It describes a comprehensive theory of how the brain works. It describes what intelligence is and how your brain creates it. The theory I present is not a completely new one. Many of the individual ideas you are about to read have existed in some form or another before, but not together in a coherent fashion. This should be expected. It is said that "new ideas" are often old ideas repackaged and reinterpreted. That certainly applies to the theory proposed here, but packaging and interpretation can make a world of difference, the difference between a mass of details and a satisfying theory. I hope it strikes you the way it does many people. A typical reaction I hear is, "It makes sense. I wouldn't have thought of intelligence this way, but now that you describe it to me I can see how it all fits together." With this knowledge most people start to see themselves a little differently. You start to observe your own behavior saying, "I understand what just happened in my head." Hopefully when you have finished this book, you will have new insight into why you think what you think and why you behave the way you behave. I also hope that some readers will be inspired to focus their careers on building intelligent machines based on the principles outlined in these pages. ...

Weren't neural networks supposed to lead to intelligent machines?
Of course the brain is made from a network of neurons, but without first understanding what the brain does, simple neural networks will be no more successful at creating intelligent machines than computer programs have been.

Why has it been so hard to figure out how the brain works?
Most scientists say that because the brain is so complicated, it will take a very long time for us to understand it. I disagree. Complexity is a symptom of confusion, not a cause. Instead, I argue we have a few intuitive but incorrect assumptions that mislead us. The biggest mistake is the belief that intelligence is defined by intelligent behavior.

What is intelligence if it isn't defined by behavior?
The brain uses vast amounts of memory to create a model of the world. Everything you know and have learned is stored in this model. The brain uses this memory-based model to make continuous predictions of future events. It is the ability to make predictions about the future that is the crux of intelligence. I will describe the brain's predictive ability in depth; it is the core idea in the book.

How does the brain work?
The seat of intelligence is the neocortex. Even though it has a great number of abilities and powerful flexibility, the neocortex is surprisingly regular in its structural details. The different parts of the neocortex, whether they are responsible for vision, hearing, touch, or language, all work on the same principles. The key to understanding the neocortex is understanding these common principles and, in particular, its hierarchical structure. We will examine the neocortex in sufficient detail to show how its structure captures the structure of the world. This will b