Learning or Cloning

Cognitive requires learning. At very least an intelligent system needs to be updated as its context changes. There are many kinds of learning.  Passive, active, statistical, inductive, example based … If I buy a book and put it on my shelf, have I learned? I am connected to the Web, so as it changes, am I more knowledgeable?

Do Cognitive systems know what they don't know?

Do I know what other people in my network know? How much effort will it take to update my cognitive system to keep it up to date? How do I know it’s up to date?

Those two last questions were major breaking points for our own use of advisory cognitive systems.

Alternatively, you could learn from cloning (or at least mimicking) your best decision makers, assuming they would know where to get the best and most current data needed. An 'expert system' style learning method. Using biological methods at their grandest. Which was recently discussed here.