Saturday, June 25, 2011

the defining data structure

The human brain is a body organ of countless cells and regions, whose chemical/electrical work consumes considerable energy to affect microscopic matter and thereby think. But pictured within a more abstract realm of information, its efforts accomplish the arrangement of data into a workable structure. It redirects, recombines, and generally tames a storm of data in the form of nerve impulses. Crudely put, brains are the coin-sorters that efficiently separate the "coins" of data by biological value. Given that a datum is never alone, i.e. information can only be informative in concert with other information, then the brain acts as the defining data structure.

Moreover, the data structure has several parts and processing steps. First, each bit of data is accompanied by its contemporary bits, into a structure of context. Second, this larger hunk of data is classified by arousing similarities to stored past data, into a structure of remembrance. Third, the data is correlated with goals and desires, into a structure of significance. Fourth, it becomes part of plans and scenarios, into a structure of proposed actions.

Since each of these tasks of definition are distinct to some extent, and also segregated by type of sensory stimuli, it's not too surprising that the brain's structure is closer to a conglomeration of specialists than a single thinly-spread generalist. When a creature develops a sense, it's cost-effective to gain narrow advantages from it in the service of common situations, e.g. fight or flight. On a limited food budget, brain cells that are well adapted to extracting information from sense data are preferable. By contrast, unshaped cells that are supremely flexible and unassuming are also not immediately helpful. Evolution prioritizes richer data structures of perception. It's more vital to absorb and distinguish raw data than to perform an orderly sequence of advanced manipulations and transformations.

Therefore, as a mirror of the brain's layout, general or abstract data structures are likely built from collections of specialty data structures. Thoughts about visual sensations, even if subconscious, happen via the brain sections that evolved to handle vision. Singers who recognize that a note's pitch is too low are using auditory specializations. Episodic memory, encompassing several senses at once, employs a wide-ranging slice of the brain. Words like "determination" or "below" or "nation-state" have connections that are more diffuse but nevertheless existent. Concepts rely on foundations, although the path can be long and twisting.

As complex as the entire mixture is, further complications apply. Instant initial sensations are not the sole data source that demands structure. For instance, incoming data can be persistent after the time of introduction, recurring by displacing recent items or fading to an activation level near sub-conscious. Either way, the long-lasting data structures have the possibility of meshing with seemingly unrelated structures to form unanticipated syntheses.

Another less traceable data source is unprovoked cell activity. Fully awake brains encountering sedate environments are nevertheless busy. Resting brains during a REM cycle are quite energetic, too. Clearly a measure of unpredictability and autonomy is built into the human organism.

Fortunately, mechanisms counteract the inherent chaos and competitiveness. Under normal circumstances data structures require consensus in order to continue having effect. No separate all-knowing judge is necessary to arbitrate. Matching data structures mutually reinforce and silence deviants, just like fitter organisms multiply and crowd out the less-fit in the absence of an overseer. A reasonable structure of data is strengthened through like-minded alliances with the expectations that it meets. A nonsensical structure of data is weakened because of its isolation.

Overall this bottom-up strategy is a gradient for progressively pushing data in the direction of verisimilitude. Inappropriate answers are often a mere starting point. Some variations on that start will link up with strong data structures in the brain, but some will not. Useful creativity is a trial-and-error enterprise that strains known constraints without breaking. Repetitive feedback establishes which prospective differences are a step forward or backward.

Hence the momentary restructuring experiments are anchored by relatively pivotal and dominant data structures. One category is urgent external reality. The freshest streams of experience are naturally vivid. Related data structures propagate by channeling the vibrancy. Remote data structures, like reveries, tend to dissipate quickly. Abnormal breaks in the division, triggered particularly by interference of ingested chemicals, can certainly reverse the usual hierarchy and produce disorienting confusion of whether data is "imaginary".

Besides urgent external reality, data structures may take a controlling role by belonging to the broad category of organism objectives. The power of basic evolutionary drives is sufficient to train a multitude of secondary aims and behaviors, perhaps resetting the original relative priorities. Emotive attraction or revulsion enriches corresponding data structures. Data that's been tied to such reactions is more likely to retain its structure. Marketers, politicians, and public-relations experts confirm the effectiveness of the tactic.

Thus the human brain begins with preexisting data structures for survival. Initially other creatures' brains are set up differently in as many ways as the creatures differ: less social creatures have lower proportions of data structures about empathy, creatures with advanced noses have more data structures devoted to recognizing smells, e.g. to discover claimed territory. Compared to the prehistoric time during which evolution encoded blunt instructions into the brain structures, not only into the human species but into its forebears, large-scale organized civilizations are the latecomers. While there isn't a biological way to transmit specific memories to offspring, inborn behavioral adaptations of general value in survival or reproduction are excellent traits for natural selection.

Sophisticated data structures produced and transmitted in human society are much more intricate and fine-grained than evolution's. However, one doesn't replace the other. The primal data structures are indispensable raw material and tools. Abstract structures can reuse earthy structures. In order to "understand" an abstaction, a concrete metaphor allows confident contemplation. Human brains are quite astute at extrapolating movement of an object; once the sequential levels of a quantity are visualized, the next level is the position the quantity would be if it were a moving object. AI data structures start from nothing. Defining data structures in the human brain start from an array of hard-won knowledge which meets a minimum standard of relevancy. Computers must be told exactly how to perform a task. Humans can be told a problem, and then they can twist their abundant data structures to define the problem and identify linked solutions.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.