Adventure Capitalist

I’m a geek. I have a ridiculous obsession with science across the spectrum: in private, I love exploring new languages, understanding linguistics, human psychology and the links between the subjects.
Professionally I have spent a fair bit of time working with machine learning (“Artificial Intelligence”), text mining and natural language processing. In between my private and professional obsessions, I have discovered some startling links that are hard to dismiss.

Are we just an n-tree of Bayesian-filter nodes?
To give a bit of background to the sub-headline: A Bayesian filter is basically how your e-mail spam-filter works - it has been taught to recognize, or rather categorize text into spam and non-spam by building what is called a “learning corpus”, a repository of background material that it knows for certain is spam and not-spam. It looks for similarities in the text, and the more background learning it has, the better the filter gets at categorizing information. These sort of filters can be taught to categorize just about any information into any type of categories: it is all based on how you teach them.

If you do not have enough background material though, the filters will be more likely to draw the wrong, sometimes irrational conclusions.
Contrast this to small children: most of our phobias and fears as grown-ups stem from mis-associations made as small children that where made because our “learning corpus”/background knowledge at the time was not comprehensive enough to distinguish irrational from rational associations. Consider for example a person with a phobia of brushing his teeth as an adult, this may stem from a painful dentist experience as a child. Obviously having a single painful experience at the dentist isn’t rational to connect to brushing your teeth, but nonetheless it happens.
It makes you wonder though, are our brains little more than a chain of Bayesian filters/categorizers?

What is being “alive? Are we just machines?
Considering that a lot of our knowledge, fears and behaviours can be explained and replicated in binary form by computers, it begs the question, at what point does sentience, self-awareness begin? If an animal, like a rabbit isn’t self-aware, is it any different in terms of “being alive” from say a tree?
Are we little more than automated machines? If a lot of our behaviour can be distilled down into more or less deterministic, binary relationships, does free will exist or is it just an illusion?

I don’t really have any answers - but starting to think about it opens a lot of interesting questions: I have certainly seen people make irrational stupid decisions, as well as seen computers make what seemed like completely autonomous, informed decisions. In the latter case, it is almost a scary realisation when your simplistic algorithms suddenly take on a life on their own and start to behave in ways that seem well beyond their sophistication.
Where does sentience begin? Or is it all just an illusion?