W3C home > Mailing lists > Public > www-rdf-logic@w3.org > November 2000

RE: Upper Cyc Ontology as MS Access DB

From: Atanas Kiryakov <naso@sirma.bg>
Date: Thu, 2 Nov 2000 19:52:45 +0200
To: <www-rdf-logic@w3.org>
Message-ID: <DEECKILKBHGOLKHEDNCDOEDMCDAA.naso@sirma.bg>
[if you didn't received the first message - it is quoted at the end]
[we apologise if you are get more then a single copy of this message]

Dear ontologists,

there was a significant interest for this database, so, we extended this
work further.


There are two news:

- the MS Access database was extended with a mapping of the EuroWordnet Top
Ontology (ETO) into the Upper Cyc Ontology (UCO). The theoretical background
could be found at: http://www.ontotext.com/index_publications.html,
[KiryakovSimov2000b].
There is a form in the database, that allows easy navigation through the
ETO. For each top concept you can see:
-- gloss of the EWN top concept
-- the list of EWN base concept associated to it
-- the list of Cyc constants related to this EWN constant
-- the gloss of the currently selected Cyc concept
-- the type of mapping of the EWN concept into Cyc (exact, difference, gap,
qualia).
-- the CycL definitions that constitute the top concept as an extension of
the Upper Cyc Ontology
-- there is also a possibility to specify some meta properties following
Guarino and Welty's classification.
The ZIP file is less than 800KB and could be downloaded for free from:
http://www.OntoText.com/downloads/index.html#ewn_to_cyc


- on-line access to the same database at http://demo.ontotext.com. There are
few differences with the form that is in the Access database:
-- Guarino and Welty's meta-properties are missing
-- for each Cyc constant related to a EWN top concept, you can see all of
the super concepts and all the collections (classes) it belongs to.
Formally, suppose there is a constant C. Let's name the set of its direct
generalisations with GEN(C) - these are all X such that (#$genl* C X). Let's
name the set of the collections where it is included explicitly with ISA
(C) - these are all X such that (#$isa C X). So, for each C, you can see
GEN(C) and ISA(C) as well as their transitive closures with respect to
$#genl* relations. For example,
    GEN(#$Person) = {#$LegalAgent, #$Primate,
#$HumanOccupationConstructResident,
			#$Omnivore }
    GEN_TRANS(#$Person) = {#$SocialBeing, #$Mammal, #$TerrestrialOrganism,
				#$Animal, #$Organism-Whole}
    ISA(#$Person) = {#$OrganismClassificationType}
    ISA_TRANS(#$Person) = {#$ConventionalClassificationType,
#$ExistingObjectType }

At the end of this message, you will find the message announcing the first
version of the database "Upper Cyc Ontology as MS Access DB" - the most
important information about the database, few words about Upper Cyc, as
well, as some guidelines for the queries etc. could be found there.

The Licensing Agreement is at the very end of the message.
Any comments are welcome

Best regards,
Atanas
---------------------------------------------------------------
Atanas Kiryakov,          http://www.sirma.bg/ak.htm
Head of OntoText Lab.,    http://www.ontotext.com
Sirma AI, Ltd. - Artificial Intelligence Labs
Phone: (359 2) 981 23 38, http://www.sirma.bg

> -----Original Message-----
> From: Atanas Kiryakov [mailto:naso@sirma.bg]
> Sent: Wednesday, June 14, 2000 1:37 PM
> To: standard-upper-ontology@ieee.org; onto-std@KSL.Stanford.EDU
> Subject: Upper Cyc Ontology as MS Access DB
>
>
> Hello,
>
> I expect that this resource in this shape could be useful for many
> people here. The MDB file is about 600KB zipped. Please contact me if
> interested.
>
> I'm attaching the message the comes usually with the file - hope, it
> will answer most of your questions.
>
> Best regards,
> Atanas Kiryakov
>
> ------------------------------------------------
> Hello,
>
> here is the Cyc's upper-level ontology as an MS Access 97 MDB file,
> zipped. It can be freely redistributed under the License Agreement at
> the end of this message.
>
> Below follow few sentences on Cyc, it's Upper Ontology and conventions.
> Next, there are some comments about the MDB file and usage instructions.
>
> At the end, you will find the License Agreement. The content of this
> message is also avalilable as ReadMe form in the database.
>
>
> [Ontology of the Upper Cyc Ontology]
> The Upper Cyc Ontology contains the most general 3000 constants out of
> the Cyc KB as well as the relations between them. From ontological
> viewpoint, most interesting seem to be the relations #$genls and #$isa.
> The first one is generalization relation and in fact have few
> variations: #$genlAttributs, #$genlPreds, #$genlInverse and #$genlMt.
> So, (#$genl* A B) in CycL means that what is represented by A is
> generalized by what is reprezented by B.
>
> It is important to mention that in Cyc #žisa relations do not have its
> most popular meaning - the one encoded in Cyc via #$genl*. In contrast,
> (#$isa I C) means that the individual I is member of the set C. So,
> #$isa is equivalent to what's usually called INSTANCE_OF if C is a class
>
> or type.
>
> It is also important to understand that Cyc constants are not concepts.
> They are just named entities in the Cyc's KB. Some of them represent
> types, but other can represent individuals, relations, functions, etc.
> The instances of ObjectType can be considered  as types. Such are 1439
> constants.
>
> More information about what is Upper Cyc Ontology can be found in the
> Cyc Ontology Guide at http://www.cyc.com/cyc-2-1/intro-public.html
>
>
> [About MDB file]
> The CYC_TOP.MDB file attached can be opened with MS Access version 97 or
>
> later. It was automatically derived from a file that identifies itself
> like that:
> ;;; Upper CYC(R) Ontology flat-file
> ;;; Copyright Cycorp 1997.  All rights reserved.
>
> It's a relational database, that can be either queried and managed via
> MS Access. Alternatively, it can be used from other programs via ODBC,
> JDBC or other similar facilities.
>
> Inside there are few tables - the names of the tables and the columns
> are the only source of intuition about the DB structure. There are also
> a number of queries that have descriptions in their Properties
> (right-click on the object, and you'll find them). See the next section
> for more information about the queries
>
> The DB has all the information available in the flat-file with the
> following exceptions:
>
> - CycL formula starting with logical operators
> (#$not,#$and,#$or,#$implies) are omitted. In total there are 403 such
> complex formula in the flat-file
>
> - the comments are cut at the 255th character. The CR and LF are
> replaced by intervals. To see the full comments, try the CYC« Ontology
> Guide at http://www.cyc.com/cyc-2-1/toc.html
>
> - for constants with multiple #$comment statements, they were
> concatenated using "<<<COMMENTn>>>" as a separator in the string
>
> One more comment is required - expressions like (#$GroupFn #$Event) are
> CycL constants even that they are not atomic constants. They are
> represented as usual constants in the MDB.
>
>
> [Queries]
> The queries can be run via the following procedure:
> - open the MDB file (in MS Access)
> - select Query tab
> - double-click on the query
> - if the query has a parameter, the system will ask you to enter a value
>
> - enjoy the result
>
> The parameters for all the queries are not case sensitive, so you'll
> obtain the same
> results for both 'MoneyForNothing' and 'moneyfornothing'.
>
> There is also one more catch - they are always interpretted as an end of
>
> the constant name. Which means that "abc" entered as a parameter value
> will be interpretted as "*abc", where * will be matched with any
> sequence of characters. It is implemented this way, to avoid typing #$
> infront of each name. However, it has certain disadvantages, while
> asking for "LexicalItem" you will get also "NonLexicalItem".
>
> To match constants that contain a string but doesn't finish with it put
> * symbol at the end. For example, "Meeting*" will match both
> "#$MeetingSomeone" and "#$MeetingTakingPlace".
>
> Typical use cases:
> - to get the direct generalizations of a constant use the "Taxonomy
> Above C" query
> - to get all the constants related to a some constant use "All relations
>
> about C"
>
> In both cases, MS Access will ask you to enter the constant name.
>
> Each query has a description accessible via right-click, Properties.
>
>
> [License Agreement]
> The conditions are exactly the same as the Cycorp's ones (see below):
> you can freely use and redistribute the MDB file, but you have to
> acknowledge
>     Atanas Kiryakov
>     Sirma AI Ltd.
>     Hr. Botev 38A
>     Sofia 1000, Bulgaria
> or via e-mail: naso@sirma.bg
>
> Using the MDB file, you should still follow the license agreement for
> the flat-file, so, both Atanas Kiryakov and Cycorp should be
> acknowledged.
>
> The original beginning of the Cycorp's source file follows
>
> ;;; Upper CYC(R) Ontology flat-file
> ;;; Copyright Cycorp 1997.  All rights reserved.
>
> ;;;Cycorp License Agreement
>
> ;;;Cycorp is providing this material from the Cyc(tm) Upper
> ;;;Ontology at no charge, for everyone to use, including
> ;;;commercial service use and incorporation into products.
> ;;;However, it is not 'Public Domain.'  Please acknowledge
> ;;;Cycorp, 3721 Executive Center Dr., Austin, TX 78731 in
> ;;;any use or citation of this material, and request that each
> ;;;further user include a full copy of this notice as well,
> ;;;in any use or citation they make of the material.  All
> ;;;these terms equally apply to renamings and other
> ;;;logically equivalent reformulations of the material in
> ;;;any natural or formal language.  Cycorp intends to
> ;;;amend and expand the material from time to time; the
> ;;;latest version is available at  http://www.cyc.com
>
> Best regards,
> Atanas Kiryakov

---------------------------------------------------------------
Atanas Kiryakov,          http://www.sirma.bg/ak.htm
Head of OntoText Lab.,    http://www.ontotext.com
Sirma AI, Ltd. - Artificial Intelligence Labs
Phone: (359 2) 981 23 38, http://www.sirma.bg
Received on Thursday, 2 November 2000 12:53:25 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 2 March 2016 11:10:32 UTC