Home

Semantic Distance: NLP Not a Resource Sink

Following the story on Mind Mappers and other RDF comments of late, I thought this NLP slide show (PDF) should get a story. Dr. Adrian Walker offers an interesting perspective in a friendly crayon-colored format, including a critique of RDF. Source site Internet Business Logic has other offerings including an online demo.

By Mark Evans at 2006-01-22 06:58 | DSL | Meta-Programming | Semantics | Theory | XML | other blogs | 13107 reads

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Minor conceptual remark for OO programmers

If you don't care much for semantic systems, realize how they already affect you. Think back to those object-orientation books which taught you to design classes with a few predicate tests like "is-a" and "has-a." RDF and NLP simply generalize that style of thought. Choices aren't limited to predicates a compiler can imitate.

By Mark Evans at Sun, 2006-01-22 07:20 | login or register to post comments

I'm not getting this one...

John Sowa's example, from the slide show:

Clyde is an elephant,
elephant is a species
==> Clyde is a species Wrong !

is used to motivate "Why we need Natural Language, even for simple semantic tasks". This is, indeed, clearly wrong, since "is a" isn't transitive because it is coalescing at least two relationships. So, the "lightweight, executable English" model is introduced:

this-item is a member of the set this-set
============================
Clyde The Elephants

this-item is a named subset of the set this-set
=================================
The Elephants All Species Of Animals

some-item is a member of the set some-set
that-set is a named subset of the set some-superset
-------------------------------------------------------------------
that-item is a member of a named subset of that-superset

(Appologies for the spacing in the tables above.)

  1. How does the original example mean that "we need Natural Language", since natural language is the problem? I am assuming Natural Language is not natural language, but is the improvement?
  2. Does anyone find the "lightweight, executable English" model either lightweight or particularly English?
By Tommy McGuire at Wed, 2006-01-25 18:20 | login or register to post comments

You should

The book Knowledge Representation by John F. Sowa has recently been published. This directory contains some information about the book, including on-line copies of the preface and the table of contents. The on-line index contains over 600 links to web sites that contain background readings about the people and topics mentioned in the book.

By Mark Evans at Sun, 2006-01-29 03:33 | login or register to post comments