Overview:

There must be a pretty good chance that AI really will happen in the 21st Century. Factasia wants a piece of the action.

What is Intelligence? By a way of clarifying the scope of the field and identifying the particular interests of Factasia, we seek a definition of intelligence which is not homo-centric.

There are many rich connections between philosophy and AI, particularly between analytic philosophy and logical AI.

This extravagent label covers Factasia's interest in the AI project of the (next!) millenium. The world transformed by the evolution of intelligence in cyberspace.

Automation of deductive reasoning is the subdomain of AI of greatest interest to Factasia. Crystal clear problems, widespread applications.

Mathematics provides the most complex applications of deductive reasoning and is therefore a testbed for the automation of reasoning. Here we consider the transition from powerful but non-intelligent programs to intelligent mechanised mathematics.

Topography

Factasia's architecture for AI places pure logic at the centre of it all.

What is Intelligence?:

By a way of clarifying the scope of the field and identifying the particular interests of Factasia, we seek a definition of intelligence which is not homo-centric.

:

There are many rich connections between philosophy and AI, particularly between analytic philosophy and logical AI.

:

This extravagent label covers Factasia's interest in the AI project of the millenium. The world transformed by the evolution of intelligence in cyberspace.

:

Automation of deductive reasoning is the subdomain of AI of greatest interest to Factasia. Crystal clear problems, widespread applications.

:

We may think of the mechanisation of mathematics as occurring in three phases, the numeric, symbolic, and logical phases. The logical phase provides a platform for mathematical AI.

numeric The predominant area of application of computers to mathematics is in brute numeric computation. Numerical analysis, takes the drudgery out of the real number computations needed in science and engineering. Discrete mathematics involving integer arithmetic is now conspicuous for its applications in cryptography.

symbolic Early AI research on symbolic mathematics has now matured into powerful software packages which transform mathematical formulae as well as undertaking numerical computation. Capabilities such as symbolic differentiation and integration bring these tools much closer to the capabilities of human mathematicians.

logical Intelligent mathematicians are able to reason about mathematics, and mechanisation ultimately depends on building tools whose capabilities are grounded in logic. A successful integration of the power of symbolic mathematics tools with suitable proof technology would provide the basis for intelligent mathematical software.

A Topography for AI

At the core of our architecture is a formal logical "inference engine". A meld of compiler and proof technologies giving fast computation of logical truths rather than data values.

Built on the logical core, the main body of applicable mathematics with just as much pure maths as helps to oil the wheels.

Beyond the theories into the applications, targeted at engineering applications. As much automated problem solving as we know how implement within the limits of energetic engineering rather than AI breakthroughs.

We seek an environment in which, in an environment full of hard graft algorithmic problem solving, intelligent capabilities can evolve and emerge. Not by natural selection. Faster than that.

Beyond logic and mathematics, beyond deduction, into empirical science. Judgement is called for here, and trusting machines may not be appropriate.

One step further, from objective science to subjective values. This is part of what human intelligence is about, but do we need it in AI?

Emotions One step further beyond the limits of machinery?

created 1995/10/7 modified 1998/08/25