Arts and Minds

Intelligence and Adaptability in Systems

My son Nova Spivack ( http://novaspivack.typepad.com/nova_spivacks_weblog/2008/02/a-classificatio.html?cid=103805366#comments ) has brought up the subject of developing a universal classification of intelligence. It is a worthwhile effort, and one that may require a century of reflection and research. It is worth more and serious work. Others have and will attempt it as well, and agreement will be slow and hard to achieve.

One problem is that we do not yet have a workable non-universal (species- our own) description of intelligence. My own questions are — What do we mean by intelligence? What are we getting at when we measure it or write about it? Much of the literature seems to confuse intelligence with 'smarts' (see my own previous posting on this blog — Is Intelligence A Property of All Life?)

I wonder if a useful way of discussing intelligence might be to consider it as an aspect of adaptability and a part of all biological process, and extend that into inorganic systems as well.

This dumps us into the possibility that intelligence evolved out of simple primal and basic properties of inorganic and organic systems in the early universe (at least on our planet, and in it’s high form (homo-sapiens) it is merely an extension of those simpler capacities for adaptability and change.

In this context intelligence is a scale of what can change or adapt in any examined system, ecosystem or species and how rapidly (in a comparative sense) this takes place. For instance, what is the scope and depth of possible change and adaptability in a molecule or virus and how does this scale up as systems become more complex? What terminology might we use to consider all this in a fresh perspective and to avoid the language and conceptual pitfalls hidden within our classical and current definitions and research?

Peer-Reviewed Publication, Weblogs and Technorati Ratings of Original Thought

This posting is another in what I now realize will be a longer series on the life-cycle and utility of communication channels. The first, posted on December 14, 2003 is entitled: Six Stages In The Life Cycle Of Communication Channels.Six Stages In The Life Cycle Of Communcation Channels Now in this current paper I will consider the special case of information propagation and dissemination for original, disruptive, or counterintuitive intellectual content.

The peer-review process filters undesirable qualities from publications within scientific and academic communities. It is generally intolerant of innovations, disruptive observations, and contributors whose work is nearly entirely original (with the exception of mathematics), yet these qualities are essential to a healthy intellectual environment.

Original workers take great risks, often remain isolated from their peers, and are typically shunned and disrespected by potential employers. They are lonely thinkers that crave colleagues and dialogue.

The web-log, or blog, is now the most accessible as well as the most rapid route to publication for these original minds, and it does offer some dialogue. But the blogosphere is a generally a chaotic and unreliable marketplace for information. It is more often used for agglomerating news, publishing news and commentary or accessing news, either personal or news of interest to the greater community, than as a portal for serious intellectual publication.

Publishing original material on a blog is risky because the contribution is automatically branded unreliable because the writers become known by the company that they keep, and that company is far too often intellectually messy and unreliable.

Artificial Intelligence Requires Emotional Mediation

Artificial intelligence will never be intelligent in the human sense until we find a way of organizing machine process, storage and retrieval that is mediated by an emotion-emulating algorithm.

Information moving about within our brains, even what we believe to be pure logical thought and fact is attached to emotional preferences and dislikes (intellectual passions if you like), and these emotional tags or neurological links assist us in making efficient and meaningful use of the primary sensory chaos present in the unprocessed perceived environment. Emotion plus other data equals meaning, and meaning is everything in both thought and emotion, and in action or communication.

Comprehending what meaning means aught to be our main target as we pursue the grail of artificial intelligence. We can eventually understand and build an operational concept of meaning, but it will be difficult (or maybe impossible) if we only stick to the computer science worldview. The difficulty will be somewhat eased if computer scientists go to lunch with psychotherapists who teach and use psychodynamic theory.

In a simple mind experiment, think of an idea or a theory, perhaps some fact or strong belief you have been working with. Are you fond of it? Do you defend this belief or theory when colleagues challenge it’s validity in meetings? Your defense is not purely logical. It is also strongly emotional.

We are motivated by emotion first, logic second. We store away and remember our observations and scientific ideas with ‘tags’ that connect emotion to logical thought.

If computation is ever to be deeply companionable with humans, we must build computers that process data the way humans feel and think, this is not improbable. Because they exist together within the brain, emotions, logically, must be merely another quality or kind of information in the brain in the same way as are logical propositions.

Emotion is not a halo of irrational spiritual vapor hovering outside our brains. It is more likely central to the brain’s own deep logic. Perhaps emotion is a faster pathway to learning and remembering in animals, including humans, and will eventually provide the same functions within computer systems and their application programs.

Perhaps, if we keep our minds open this avenue of investigation may also lead to a better understanding of the mysterious process of human thought and emotion.

Information Sausages

Imagine that the worldwide network is in fact only a information sausage exchange and sausage packing plant . If we poke a peephole in the roof and look down at the operations below, we see crowds of people, millions really, trying to stuff their own information into some sausage and send it off, or pick-up a delivery of sausage with their name on it. Each person arrives with some idea or question to stuff into the sausage-making machines below. We see that the production line winds around like an airport check-in area—unrelated people are located in front of and behind each other, each in their turn stuffing their information into funnels, filling each of the sausage-skins in sequence with discontinuous, unrelated packages of information.

The information-sausages move along the line, each filled with it’s bits, and are cut off from the endless supply at the end of the line where a packing station counts lengths of just-so-many-sausages to be randomly tossed into boxes and shipped out the back door. From our perch on the roof we see outside the building that the boxes are carried away and distributed through the distribution network.

At millions of endpoints and nodes in this network, like the one you are on right now, humans get to sample the sausage and digest it’s contents. But the overall impression we get from looking down through our peephole is that of too many people trying to jam too much stuff into too many small packages and tossing them unsorted, into an endless queue of trucks. We are looking at a traffic jam stretching from input to destination of ideas, words, bits, identities, locations, and workers, each speaking different languages, without understanding of meaning. It is Babel, even for those who speak the same language. The hum is deafening, the noise out-shouts the signal. This signal to noise imbalance is most difficult when nasty selfish folks attempt to fill millions of sausage links with viruses causing endless trouble. While the sum of all this effort is greater than the sausages themselves, it is not as great as it should be. We have a thoughtless network because information does not conform to semantic structures. We need a thoughtful one structured in the terms of human language. We need Twine (developed by Radar Networks Inc.) to tie our packages together in personalized 'giftwrap'.

After The Fast Last Mile are The Slow Last Inches

After the ‘last mile’ is completed, when the last high bandwidth cable has been connected, and computers are predictably faster than they are, information will still have to travel the last few inches from the screen and be formed into meaning and memory within the mind. These inches are your own nerve fiber, not copper or optical fiber, and they place the ultimate limits upon our efforts to push or pull information from providers to consumers. Outwitting our own brains will be the next big thing. In order to make these last inches more receptive to what the information network provides to their computer screens, the network must pre-digest the information it serves to users. The most powerful digestive juices we can employ come by way of Twine via Radar Networks Inc.. Twine, like the human brain, does a lot of it's work associatively. It is the first really syncretic system to be developed for computation. Using Twine, my computer finds something for me, something important that I did not know and was not seeking. It is a bit of a shock. This is new territory. My computer suddenly seems smart.

While Moore’s Law optimizes the possibilities for many aspects of computation, the pace of nerve fiber information transmission will ever remain constant. Eventually, (a long time from now) a computer may demonstrate the processing and intelligence equivalent of brain-power. But we will still always have to read, organize and consider the sentences as we read them.

Then follows the complex brain-work of deriving and ascribing context and meaning to what we have read. This will always happen at the nice human speed of brain and nerve. There is human pleasure in this process and pace, like taking a walk, or a swim, we live within and enjoy our human scale speed limits. Someday Moore’s Law will become unimportant to most hands-on uses of computers because our brains will be so much slower (and more expensive).

The Twain Shall Meet At Last!

It has been a guiding principle in our learning, drummed into our brains during primary school years, that one cannot compare apples and oranges because they are different (despite similarities obvious to any schoolchild), and that ‘never the twain shall meet’— that because of these differences they can never be usefully compared or combined. Soon we will all know that it has always been a lie. The twain shall be tied together by Twine.

Today, within minutes after I received my notice that Twine was being demoed At the Web 2.0 conference on Friday, October 19, ’07, I entered the search terms (Twine, radar,) into Google. What I got were eight references to today’s Radar Network’s announcement of the pioneering product Twine, along with an overwhelming number of references to all sorts of things I don’t want to bother with from nubs of string to space-aliens. Google brought me far too many irrelevant pages-full-of-pages, signifying nothing. Google regurgitated the whole hairball including some few useful threads that were not always up front, or even within the first few pages.

That problem, and others, have been addressed by the new product, ‘Twine’, developed by Nova Spivack and his team at Radar Networks. Twine will accomplish at least three grand feats.