NOTE: Generality In Artificial Intelligence

(This blog will be written in my shabby English, sorry for that in advance...because otherwise, it will become a mixture of Chinese explanations with English citations...)


1. John L. McCarthy. Generality in artificial intelligence. Commun. ACM, 30(12):1029--1035, 1987.


It is really great to look back into the classical papers, and 1 is one of them.

Well, I know something but they are in a mess. In this paper, the knowledge pieces in my mind are tied and positioned to the right place. And, by looking back we can tell what are the basic problems and which of them have been solved, if not, what are the contributions. Most importantly, Issues discussed in the Semantic Web community can also find their originals here.

About the Semantic web. We know stanford KSL lab has contributed a lot (e.g. ref. the work of Dr. Thomas R. Gruber and Dr. Deborah L. McGuinness ), where Prof. McCarthy worked for 40years.
Some ideas explained in these paper are still being explored by SW people.

Make some notes as follows.
-----------------------------
1.Horn clause & prolog
-----------------------------
from wikipeida
from StarJumper

-----------------------------
2. Situation calculus.
-----------------------------
Have met them before in PKU, in the lecture given by Prof. Zhuoqun Xu on Artificial Intelligence.Got to know it from the the book: A modern approach, artificial intelligence. And now I know why it is proposed, and what does it mean for AI.


-----------------------------
3. Qualification problem &
frame problem.
-----------------------------
-> circumscription.
[todo]

-----------------------------
4. Reification
-----------------------------
reification is a mechanism provided by RDF. Also there is a long discussion about it in the mailing list of rdf and topic map.
Why reification is important? In RDF is means that one can say something about the predicate. And in 1, the section of Reification begins with:
"Reasoning about knowledge, beliefs, or goals requires extensions of the domain of objects being reasoned about..."
and in the example, "and predicate constant.... is taken as an object in the first order language.."


-----------------------------
5.Notion of content
-----------------------------
AI people is always facing with the "close world problem": "Whenever we write an axiom , a critic can say it is true only in a certain context. With a little ingenuity, the critic can usually devise a more general context in which the precise form of the axiom does not hold. Looking at human reasoning as reflected in language emphasizes this point."

And a proposed possible "way out involves formalizing the notion of context an combining it with the circumscription method of nonmonotonic reasoning."

The idea of context is: a sentence is meaningful in particular context.
With the conception that ontology is an "agreed specification", I redirected my focus to context and ontology.
There are a lot of related work already(C&O, 2, 3 ), and seems it is attracting more and more attention from SW.

[todo]Problems and Projections in CS for the Next 49 Years



-------
2. Aberer, K., Cudre-Mauroux, P., Ouksel, A.M., Catarci, T., Hacid, M.S. Illarramendi, A., Kashyap, V., Mecella, M., Mena, E., Neuhold, E.J., Troyer, O.D., Risse, T., Scannapieco, M., Saltor, F., De Santis, L., Spaccapietra, S., Staab, S., Studer, R. (2004). Emergent Semantics Principles and Issues. Database Systems for Advanced Applications 9th International Conference, DASFAA.

3.Mika, P. (2005). Ontologies Are Us: A Unified Model of Social Networks and Semantics. International Semantic Web Conference.