Re: Commercial video game that learns from its own mistakes?

"baylor" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]
> So, the point being, an ANN starts in learning mode, at which point it
> doesn't actually do anything but learn (i.e., it doesn't make
> decisions) and then ends up in usability mode, at which point it has
> learned all it will ever learn and never adjusts again

Which brings up a point: Is it possible for such networks to continue
learning? Does the word Traditional apply to these algorithms? Is the
decision to cut off learning a necessary threshold point? I'm wondering if
perhaps the off-line learning model was created to conserve cpu cycles, a
resource that has climbed by orders of magnitude over the past decade.

I seem to recall that Learning is opposite Memorization; a balance between
flexibility and rigid precision. What if we let the system continue to
learn? It is my guess that such a system would 'forget' old information as
new information continued to update the weights, which seems to mirror
nature.