No Distinction between human and artificial intell
Making No Distinction Between Human and Artificial Intelligence
This assumption was used to constrain the final structure of the Soar
architecture. As mentioned in the discussion on the focus on the
cognitive band, there is an abundance of
psychological data that could be used to fix performance universals.
For example, the architecture was used to generate a model of the
power law of practice which in turn led to the inclusion of
chunking as Soar's
general learning mechanism. There are several other benefits to this
approach as well:
- Artificial experiments (applying the architecture to general problems
requiring intelligence) results in a measure of the architecture's
general sufficiency which can be compared to human sufficiency in similar
situations;
- Theoretical conclusions force a consideration of the general
requirements of intelligence;
- Comparative analysis of human performance and Soar systems results in
feedback concerning how new architectural mechanisms may be
appropriate for Soar.
To return, press HOME.
To go to the next document, press NEXT.