From: Xu Miao (xm_at_u.washington.edu)
Date: Wed Oct 22 2003 - 11:11:48 PDT
Title: The Theses of Knowledge Representation
Author: Jon Dyle, Ramesh S. Patil
Summary:
This paper demonstrate argue that restricted knowledge representation system
limits the general-purpose faciities by a series of examples.
The most important ideas in the paper are:
1. Due to the restrict proposed by Levesque and Branchman, some definition
or concepts are inexpressible. He gave some very detailed examples showing
how it works. The restrict on ommit construct also reduce the utility of the
classification And the require of quickness misconceives the notion of
runtime efficiency.
2. The soundness, comleteness and worst-case complexity are inadequate
measures for evaluating the utility.
The flaw of the paper:
Actually it is observable that restrict on KB presentation will cause the
reduction of the ability of representation. Author showed in very details
about this idea, but didn't give enough explanation and solution. His argue
only demonstrates another point of view of the problem, not prove that
Levesque and Brachman were false on their argument.
A possible research questions:
1. Using statistic theory measuring the efficiency.
2. Tolerate incompleteness or complete.
This archive was generated by hypermail 2.1.6 : Wed Oct 22 2003 - 11:11:24 PDT