[comp.ai] Wanted: Tools and references on Frames & Semantic Nets

heintze@grane.enet.dec.com (Siegfried Heintze) (09/10/90)

I would like some examples of using frames or semantic nets.  I understand these
two terms
are synonymous.
 
   (1)  Can anyone reference me to a text book example I can try out to better
understand
          how to use frames?   I have "Knowledge Systems and Prolog" By Walker
(et al) which 
          contains a section on frames,  but the examples are trivial and
incomplete.

   (2)   What tools are available for implementing frames and semantic nets? 
Apparently
          Prolog is one option.  I've heard  that ART, Knowledge Craft, KEE are
options.  

          (a) What other options are there.
          (b)  Can anybody give me address and phone numbers for ART, Knowledge
craft and KEE
                 so I can aquire some literature from them?


                              Sieg

asa@vall.dsv.su.se (]sa Rudstr|m) (09/11/90)

In article <15260@shlump.nac.dec.com> heintze@grane.enet.dec.com (Siegfried Heintze) writes:
>I would like some examples of using frames or semantic nets.  I understand 
>these two terms are synonymous.
> 
> ...

WHAT!!! I do not agree at all that "frames" and "semantic nets" are
synonymous. If you open any introductory book on AI, you will find that
most (all?) authors agree with me. Both are structures for representing
knowledge, but...

Originally, a frame was a general term denoting a description of some kind
of stereotypical situation ("... like being at a child's birthday party...", 
see "A framework for representing knowledge" by Marvin Minsky, in Winston:
The Psychology of Computer Vision,McGraw-Hill, 1975 (Minsky is the one who
"invented" the term)). However, the term has now come to denote a kind of 
record structure, with some extra features (inheritance,procedural 
attachment...). 

KEE, for example, uses frames for representing both objects and rules.

A semantic network is a more fuzzy term, but one example is John Sowa's
conceptual structures (Addison-Wesley, 1984)


Aasa Rudstrom

dclark@ncratl.Atlanta.NCR.COM (Dave Clark) (09/11/90)

The Luger & Stubblefield text is a very good introduction to frames and
semantic nets although its examples may not be as deep as what you're
looking for.

   Luger, George F. and Stubblefield, William A., Artificial Intelligence
   and the Design of Expert Systems, Redwood City: Benjamin Cummings,
   1989, ISBN 0-8053-0139-9.

It has an excellent collection of references for several AI areas.

By the way, although frames and semantic nets are both used for knowledge
representation, they are not exactly the same thing.  Semantic nets
organize knowledge using explicit links between objects in the knowledge
base.  Frames (also, schemas) are single, complex entities that represent
complex situations or objects.  Frames closely resemble records as in
Pascal, and are intimately related to object-oriented systems.

-- 
Dave Clark                      (404) 623-7367
NCR E&M Atlanta                 VOICEplus 751-7367
2651 Satellite Blvd.            dclark@ncratl.Atlanta.NCR.COM
Duluth, GA  30136               (setf (get 'dave 'disclaimer) 'standard)

kumard@acsu.buffalo.edu (Deepak Kumar) (09/12/90)

See papers/articles on SNePS- Semantic Network Processing System.

For a detailed bibliography of SNePS papers and information on
availability send mail to : snwiz@cs.buffalo.edu

See

SNePS Considered an Intensional Propositional Semantic Network
Shapiro & Rapaport
in McCalla and Cercone (Eds)
THE KNOWLEDGE FRONTIER

also

Current Trends in SNePS 
Volume 437, Lecture Notes in AI
D. Kumar (ed)
Springer Verlag, 1990

Deepak.


-- 
"I think I know why the dog howls at the moon" - JC
kumard@cs.buffalo.EDU
kumard%cs.buffalo.edu@ubvm.bitnet
Deepak Kumar, Dept. of CS, 226 Bell Hall, SUNY@Buffalo, NY 14260.

rapp@balr.com (Chuck Rapp) (09/12/90)

Aasa brings up an interesting question: does there exist a theory
regarding representational power?  In computational theory, two
different forms of computational are said to have equivalent power
if all computations in form one can be done in form two and vice
versa.  Has anyone studied the equivalence in representational power
between semantic nets, frames, and logic or between declarative
representation and procedural?  Is representational power another
form of computational power?

While there are differences between semantic networks and frames, if
they have the same representational power, then the only reason to
choose between them is ease of use in representing what you want.
-- 
+---------------+-------------------+-------------------------------+
| Chuck Rapp    |  rapp@balr.com    | Also: rapp@tellabs.com        |
| BALR, Inc.    |  uunet!balr!rapp  |                               |
+---------------+-------------------+-------------------------------+