The Komogorov complexity is what the neocortex does, or what autoencoders do, while the statistical approaches are a fallback mechanism that gets used when the neocortex can't build a model of something. It gets extra confusing, because you can be "autistic" both because you rely on the neocortex significantly more or significantly less than the general population.

As for the examples:

The stronger the neocortex, the better and more flexibly language can be used to say precisely what is meant. The weaker it gets, the more rigid the language. A large, rigid vocabulary is needed to communicate anything. The ability to derive words may disappear, in languages that allow it. Memorized phrases can take arbitrary meanings, words accumulate random "collocations" from the imprecise learning, and metaphors may be a fallback mechanism to replace words that can't be understood easily, or are not available in the language, or not known. This may cause confusion in more neocortex based speakers of the superficially identical language.

The stronger the neocortex, the less needs to be explicitly taught, as the person can build models that are good enough on their own.

The stronger the neocortex, the more can be seen (and heard), because you deal with better packed information, with much smaller volumes than more poorly packed, or even raw data. This likely also allows you to remember more, as there is less to remember.

The neocortex use appears to be negatively correlated with IQ.

Posts

Wikitag Contributions

Comments