Tag Archives: AI

Storing Errors in the Brain and Creative Sparks

It occurs to me that there could be a connection between those mistakes we make, which we store as truths for even small fractions of time, and the later ability to create or innovate.  As I can at this time conceive of no scientific proof by which this speculation may be either confirmed or denied, it must remain solidly in the realm of philosophy.  First I will outline what I mean and then I will look at the possible consequences should this be true.

The brain stores memories, experiences, and the like as synaptical connections within its neural networks.  Some of these stores data points we can think of as true and others false.  Consider a child who encounters a small furry four-legged animal.  They may, adorably, call that animal a dog.  This experience will become part of their neural matrix relating to dogs and fur and animals &c—complete with many storage points, interrelations, and connections.  When later that child discovers this animal was and is in fact a badger, adjustments will be made across the matrix to allow the child to correctly distinguish between the true dog and the true badger.  However, what is important here is the false badger.

For the duration in which the badger was falsely identified as a dog, pathways were formed and connections were created all of which can be called upon later.  Clearly the person in question will not want to falsely identify the badger as a dog, but these pathways and connections provide alternative routes of thinking along which new ideas and innovations might be prospected.

Consider next that any brain, young or old, will contain thousands upon thousands of these false truths.  They may have been briefly held or long-standing, what matters is that what is false was thought true.  This branching allows for later branching.  It seems to me that this later branching can assist in the creative and innovative processes—branching into new ideas heretofore unthought.

I do not claim that this is the only nor the necessary cause of the creative spark, merely that it can be considered as a partner in the innovative process.

If this is the case, then we might surmise that any brain which was not capable of allowing for false truths of some minimal duration would also be in capable of exploiting those alternate pathways for creation and innovation.  Clearly this would have an impact in the field of Artificial Intelligence.  Such an artificial brain must be able to store indefinitely and use at least occasionally information which is incorrect.

My understanding of the current breed of AI decision making networks is that they are able to learn by adjusting weights over time.  These weights do not store the incorrect pathways; rather, they replace them.  This may well yield positive results in creating brain-like computing devices; however, it may be necessary to allow these artificial brains to maintain databases of false truths, mistakes, and incorrect impulses if we are to see them create and innovate.

This is speculative and probably highly condemnable by stronger philosophers than myself, but here it is anyway, out in the theater of ideas ripe for your consideration.