Abstract
In an awesome review of AI John Launchbury, special assistant to DIRO, DARPA, defined four dimensions of processing information:
- perceiving
- learning
- abstracting
- reasoning
This is not any perspective on AI, it is a perspective from the founders and pioneers of internet. Although there has been significant progress with first and second generation AI systems in reasoning, learning and perceiving, abstraction has not been part of the game. In fact Launchbury could not say much about this in his talk.
Nevertheless he explained clearly where we are heading in the third generation of AI systems. Models and contextual information will play a critical role. But this is a pretty close match with Aristotle’s theory of semiosis. The mechanism of abstraction is related to the ideal world of models and the world of real objects. Those two worlds are bridged by symbols, signs and these create meaning in contextual information. So this is the objective of third generation AI.
It is not a coincidence that I have been investigating since 2012 how this triangle of meaning can be applied to the problem of information representation. Recently I have presented officially to the public R3DM/S3DM data modeling framework in European Wolfram Technology Conference. In one of his talks Conrad Wolfram emphasized how important is abstraction to education and learning. But so far this mechanism of abstraction has not been computable. This is what we anticipate to see in the near future.
The mechanism of abstraction can unify these other three processes of perception-interpretation, learning, and reasoning. And it can also interconnect every thing on the internet, every bit of information, in the very same way our computers are interconnected with IP addresses.
We are here to build powerful, meaningful relationships easily. This is our mantra after all.