OWL vs. FOL

classic Classic list List threaded Threaded
9 messages Options
Reply | Threaded
Open this post in threaded view
|

OWL vs. FOL

Daniel Gross-2
Hi, 

I have some experience with Protege 3.x, the frame based ontology building, am introducing myself (again) to OWL (and later OWL2). 

I am very interested in learning more about the expressiveness of OWL, OWL2 and various Description Logics (by example), as well as better understanding in what way OWL/OWL2 and DLs are less expressive than FOL.

I am wondering:

a. Is there some introductory text by example that presents the topic of expressivness of the many DL variants and in relation to FOL
b. Is there perhaps a protege "chat room" (on skype, irc, elsehwere) where I can chat about this (and other topics)

On a different note:

I also started to look at Bayesian Belief Network approach to represent causal relationships between evidences and conclusions. I also noticed works that seem to connect OWL with BBNs. How mature is such combination considered; are there any known implementations with which there exists some experience. 


thank you,

Daniel
 

--

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: OWL vs. FOL

carter01

Daniel,

 

It is important to understand the differences in the terms you are using. The foundations of ontology are grounded in the metaphysical attempt to understand the nature of the “thing”. First Order Logic (FOL) is the formal mathematical system using quantified variables to define objects or the “thing”. FOL (also known as predicate logic or calculus) is distinguished from propositional logic that does not use quantified variables. Since FOL uses quantified variable, it allows for a higher degree of flexibility to describe the nature of a “thing” in essence by declaration.

 

Description Logic (DL) is a family of formal knowledge representation and as such has specific notations used for description. It is more expressive than propositional logic and more efficient than FOL due to specific operators and notations used in each individual language. The DL languages include: Attributive (AL), Frame-based (FL), and Existential (EL) that are followed by extension modifiers that define restrictions. There are several patterns that can be used in DL that are common such as, SHIQ, ALCOIN, SHOIN, SHIF, etc… Each variation of DL is formed by the annotations of the base logic followed by the restrictions used in the logic. OWL2 ontologies often expresses the “SRIQ” logic (S being the annotation for ALC or Attributive Language Complex concept negation, followed by R- limited complex role inclusion, I- Inverse properties, Q- Qualified cardinality restriction. FOL and DL can mathematically define a set of relationships but they are difficult to instantiate in a computer processing system.

 

The Web Ontology Language (OWL) is a computational markup language for expressing the complex relationships about “things” and groups of “things” within the context of the semantic web. From an IT perspective, an ontology is a specification of a conceptualization (Gruber). OWL2 provides the language for computability. A key part of developing an ontology (in OWL or any other framework) is to provide the logic that forms the relationships in a document called a specification. The logic can be annotated in FOL or DL but it must be annotated such that its explicit. There are many examples that can be used to define the process (the infamous Pizza Ontology for example). There are resource aplenty but they do take time to read, digest, and integrate into your mindscape. Here is my list, although may have some or most of these already:

 

Semantic Web

http://www.w3.org/standards/semanticweb/

http://searchsoa.techtarget.com/tip/Reviewing-the-elements-of-the-Semantic-Web

 

OWL

Note: It is important to remember that there are different grammar profiles that offer specific characteristics both normative and non-normative.

http://www.w3.org/TR/#tr_OWL_Web_Ontology_Language

http://semanticweb.org/wiki/OWL_2

http://www.w3.org/TR/owl2-profiles/#Appendix:_Complete_Grammars_for_Profiles

http://webont.org/owled/2008/papers/owled2008eu_submission_23.pdf - Excellent paper for describing how it all works

 

Resource Description Framework

http://www.w3.org/RDF/

http://www-archive.mozilla.org/rdf/doc/examples.html - excellent example of RDF use

 

Linked Data

http://www.w3.org/TR/#tr_Linked_Data

 

First Order Logic

http://www.cs.cornell.edu/courses/cs4700/2011fa/lectures/16_FirstOrderLogic.pdf

http://mathworld.wolfram.com/First-OrderLogic.html

http://www.philosophypages.com/lg/index.htm

 

Descriptive Languages

https://en.wikipedia.org/wiki/Description_logic -Usually eschew Wikipedia for formal writing but this is a good comparison with excellent references.

 

I hope this helps you on your way…

 

Mark

 

 

From: protege-user [mailto:[hidden email]] On Behalf Of Daniel Gross
Sent: Monday, July 6, 2015 8:49 AM
To: [hidden email]
Subject: [protege-user] OWL vs. FOL

 

Hi, 

 

I have some experience with Protege 3.x, the frame based ontology building, am introducing myself (again) to OWL (and later OWL2). 

 

I am very interested in learning more about the expressiveness of OWL, OWL2 and various Description Logics (by example), as well as better understanding in what way OWL/OWL2 and DLs are less expressive than FOL.

 

I am wondering:

 

a. Is there some introductory text by example that presents the topic of expressivness of the many DL variants and in relation to FOL

b. Is there perhaps a protege "chat room" (on skype, irc, elsehwere) where I can chat about this (and other topics)

 

On a different note:

 

I also started to look at Bayesian Belief Network approach to represent causal relationships between evidences and conclusions. I also noticed works that seem to connect OWL with BBNs. How mature is such combination considered; are there any known implementations with which there exists some experience. 

 

 

thank you,

 

Daniel

 

 

--


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: OWL vs. FOL

Luis Enrique Ramos García
In reply to this post by Daniel Gross-2
Dear Daniel, 

for the expressiveness of DL, I recommend you  this reference:



There, some mentioned to FOL is done, but it stresses in DL family. 



Best regards


Luis Ramos

2015-07-06 9:19 GMT-04:30 Daniel Gross <[hidden email]>:
Hi, 

I have some experience with Protege 3.x, the frame based ontology building, am introducing myself (again) to OWL (and later OWL2). 

I am very interested in learning more about the expressiveness of OWL, OWL2 and various Description Logics (by example), as well as better understanding in what way OWL/OWL2 and DLs are less expressive than FOL.

I am wondering:

a. Is there some introductory text by example that presents the topic of expressivness of the many DL variants and in relation to FOL
b. Is there perhaps a protege "chat room" (on skype, irc, elsehwere) where I can chat about this (and other topics)

On a different note:

I also started to look at Bayesian Belief Network approach to represent causal relationships between evidences and conclusions. I also noticed works that seem to connect OWL with BBNs. How mature is such combination considered; are there any known implementations with which there exists some experience. 


thank you,

Daniel
 

--

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: OWL vs. FOL

Pavel Klinov
In reply to this post by Daniel Gross-2
On Mon, Jul 6, 2015 at 3:49 PM, Daniel Gross <[hidden email]> wrote:
> Hi,
>
> I have some experience with Protege 3.x, the frame based ontology building,
> am introducing myself (again) to OWL (and later OWL2).
>
> I am very interested in learning more about the expressiveness of OWL, OWL2
> and various Description Logics (by example), as well as better understanding
> in what way OWL/OWL2 and DLs are less expressive than FOL.

DLs are subsets of the two variable fragment of FOL with counting
quantifiers (should be familiar to you if you come with the FOL
background). There's a correct syntactic translation of DL formulae
into FOL sentences. It's been published in a number of papers but you
may first checkout p.25 of
http://www.aifb.kit.edu/images/1/19/DL-Intro.pdf

This means that things that you can't do in C^2, you also can't express in DLs.

This gives you the upper bound on expressiveness and complexity but
there's a bunch of logics in the DL family with less expressivity.
Some of them, e.g. EL, are formal foundations of the OWL 2 Profiles.

>
> I am wondering:
>
> a. Is there some introductory text by example that presents the topic of
> expressivness of the many DL variants and in relation to FOL

There's the DL Handbook and this: http://arxiv.org/pdf/1201.4089.pdf.
Should be enough to get you going.

> b. Is there perhaps a protege "chat room" (on skype, irc, elsehwere) where I
> can chat about this (and other topics)
>
> On a different note:
>
> I also started to look at Bayesian Belief Network approach to represent
> causal relationships between evidences and conclusions. I also noticed works
> that seem to connect OWL with BBNs. How mature is such combination
> considered; are there any known implementations with which there exists some
> experience.

Not very mature but some tools exist, see
https://github.com/julianmendez/born as a recent example.

Cheers,
Pavel

>
>
> thank you,
>
> Daniel
>
>
> --
>
> --
> Dr. Daniel Gross
> http://www.linkedin.com/in/grossd18
>
>
> _______________________________________________
> protege-user mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/protege-user
>
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: OWL vs. FOL

Pavel Klinov
In reply to this post by carter01
On Mon, Jul 6, 2015 at 5:03 PM, Mark A. Carter-Gmail
<[hidden email]> wrote:

> Daniel,
>
>
>
> It is important to understand the differences in the terms you are using.
> The foundations of ontology are grounded in the metaphysical attempt to
> understand the nature of the “thing”. First Order Logic (FOL) is the formal
> mathematical system using quantified variables to define objects or the
> “thing”. FOL (also known as predicate logic or calculus) is distinguished
> from propositional logic that does not use quantified variables. Since FOL
> uses quantified variable, it allows for a higher degree of flexibility to
> describe the nature of a “thing” in essence by declaration.
>
>
>
> Description Logic (DL) is a family of formal knowledge representation and as
> such has specific notations used for description. It is more expressive than
> propositional logic

This is strictly speaking not true. Not all DLs are more expressive
than propositional logic. All polynomial fragments, e.g., the EL
family, do not subsume propositional logic. Neither do DLs from the
DL-Lite family.

Cheers,
Pavel

> and more efficient than FOL due to specific operators
> and notations used in each individual language. The DL languages include:
> Attributive (AL), Frame-based (FL), and Existential (EL) that are followed
> by extension modifiers that define restrictions. There are several patterns
> that can be used in DL that are common such as, SHIQ, ALCOIN, SHOIN, SHIF,
> etc… Each variation of DL is formed by the annotations of the base logic
> followed by the restrictions used in the logic. OWL2 ontologies often
> expresses the “SRIQ” logic (S being the annotation for ALC or Attributive
> Language Complex concept negation, followed by R- limited complex role
> inclusion, I- Inverse properties, Q- Qualified cardinality restriction. FOL
> and DL can mathematically define a set of relationships but they are
> difficult to instantiate in a computer processing system.
>
>
>
> The Web Ontology Language (OWL) is a computational markup language for
> expressing the complex relationships about “things” and groups of “things”
> within the context of the semantic web. From an IT perspective, an ontology
> is a specification of a conceptualization (Gruber). OWL2 provides the
> language for computability. A key part of developing an ontology (in OWL or
> any other framework) is to provide the logic that forms the relationships in
> a document called a specification. The logic can be annotated in FOL or DL
> but it must be annotated such that its explicit. There are many examples
> that can be used to define the process (the infamous Pizza Ontology for
> example). There are resource aplenty but they do take time to read, digest,
> and integrate into your mindscape. Here is my list, although may have some
> or most of these already:
>
>
>
> Semantic Web
>
> http://www.w3.org/standards/semanticweb/
>
> http://searchsoa.techtarget.com/tip/Reviewing-the-elements-of-the-Semantic-Web
>
>
>
> OWL
>
> Note: It is important to remember that there are different grammar profiles
> that offer specific characteristics both normative and non-normative.
>
> http://www.w3.org/TR/#tr_OWL_Web_Ontology_Language
>
> http://semanticweb.org/wiki/OWL_2
>
> http://www.w3.org/TR/owl2-profiles/#Appendix:_Complete_Grammars_for_Profiles
>
> http://webont.org/owled/2008/papers/owled2008eu_submission_23.pdf -
> Excellent paper for describing how it all works
>
>
>
> Resource Description Framework
>
> http://www.w3.org/RDF/
>
> http://www-archive.mozilla.org/rdf/doc/examples.html - excellent example of
> RDF use
>
>
>
> Linked Data
>
> http://www.w3.org/TR/#tr_Linked_Data
>
>
>
> First Order Logic
>
> http://www.cs.cornell.edu/courses/cs4700/2011fa/lectures/16_FirstOrderLogic.pdf
>
> http://mathworld.wolfram.com/First-OrderLogic.html
>
> http://www.philosophypages.com/lg/index.htm
>
>
>
> Descriptive Languages
>
> https://en.wikipedia.org/wiki/Description_logic -Usually eschew Wikipedia
> for formal writing but this is a good comparison with excellent references.
>
>
>
> I hope this helps you on your way…
>
>
>
> Mark
>
>
>
>
>
> From: protege-user [mailto:[hidden email]] On
> Behalf Of Daniel Gross
> Sent: Monday, July 6, 2015 8:49 AM
> To: [hidden email]
> Subject: [protege-user] OWL vs. FOL
>
>
>
> Hi,
>
>
>
> I have some experience with Protege 3.x, the frame based ontology building,
> am introducing myself (again) to OWL (and later OWL2).
>
>
>
> I am very interested in learning more about the expressiveness of OWL, OWL2
> and various Description Logics (by example), as well as better understanding
> in what way OWL/OWL2 and DLs are less expressive than FOL.
>
>
>
> I am wondering:
>
>
>
> a. Is there some introductory text by example that presents the topic of
> expressivness of the many DL variants and in relation to FOL
>
> b. Is there perhaps a protege "chat room" (on skype, irc, elsehwere) where I
> can chat about this (and other topics)
>
>
>
> On a different note:
>
>
>
> I also started to look at Bayesian Belief Network approach to represent
> causal relationships between evidences and conclusions. I also noticed works
> that seem to connect OWL with BBNs. How mature is such combination
> considered; are there any known implementations with which there exists some
> experience.
>
>
>
>
>
> thank you,
>
>
>
> Daniel
>
>
>
>
>
> --
>
> --
>
> Dr. Daniel Gross
>
> http://www.linkedin.com/in/grossd18
>
>
> _______________________________________________
> protege-user mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/protege-user
>
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: OWL vs. FOL

Daniel Gross-2
In reply to this post by carter01
Dear Mark,

Many thanks for the detail response.

Perhaps I should also ask about purpose and application areas for
Description Logic (DL) knowledge representations. When describing a
conceptualization of some things (say a Pizza business) using some kinds of
DL and perhaps even embodied into KBs (such as triple stores that support
DL kind of inferences). What are the typical applications areas. Is it
about standardizing definitions of terminology, or are there also
applications areas for the creation of intelligent/AI kind of systems that
take advantage of kinds of DL reasoning support?

thank you,

Daniel

p.s. I guess variants of FOL, such as implemented in Prolog as well as in
some agent-oriented logic programming platforms, have been used for more
"general purpose" intelligence/reasoning. DLs are seen as more restricted
in their application areas?

On Mon, Jul 6, 2015 at 6:03 PM, Mark A. Carter-Gmail <[hidden email]> wrote:

Daniel,

 

It is important to understand the differences in the terms you are using. The foundations of ontology are grounded in the metaphysical attempt to understand the nature of the “thing”. First Order Logic (FOL) is the formal mathematical system using quantified variables to define objects or the “thing”. FOL (also known as predicate logic or calculus) is distinguished from propositional logic that does not use quantified variables. Since FOL uses quantified variable, it allows for a higher degree of flexibility to describe the nature of a “thing” in essence by declaration.

 

Description Logic (DL) is a family of formal knowledge representation and as such has specific notations used for description. It is more expressive than propositional logic and more efficient than FOL due to specific operators and notations used in each individual language. The DL languages include: Attributive (AL), Frame-based (FL), and Existential (EL) that are followed by extension modifiers that define restrictions. There are several patterns that can be used in DL that are common such as, SHIQ, ALCOIN, SHOIN, SHIF, etc… Each variation of DL is formed by the annotations of the base logic followed by the restrictions used in the logic. OWL2 ontologies often expresses the “SRIQ” logic (S being the annotation for ALC or Attributive Language Complex concept negation, followed by R- limited complex role inclusion, I- Inverse properties, Q- Qualified cardinality restriction. FOL and DL can mathematically define a set of relationships but they are difficult to instantiate in a computer processing system.

 

The Web Ontology Language (OWL) is a computational markup language for expressing the complex relationships about “things” and groups of “things” within the context of the semantic web. From an IT perspective, an ontology is a specification of a conceptualization (Gruber). OWL2 provides the language for computability. A key part of developing an ontology (in OWL or any other framework) is to provide the logic that forms the relationships in a document called a specification. The logic can be annotated in FOL or DL but it must be annotated such that its explicit. There are many examples that can be used to define the process (the infamous Pizza Ontology for example). There are resource aplenty but they do take time to read, digest, and integrate into your mindscape. Here is my list, although may have some or most of these already:

 

Semantic Web

http://www.w3.org/standards/semanticweb/

http://searchsoa.techtarget.com/tip/Reviewing-the-elements-of-the-Semantic-Web

 

OWL

Note: It is important to remember that there are different grammar profiles that offer specific characteristics both normative and non-normative.

http://www.w3.org/TR/#tr_OWL_Web_Ontology_Language

http://semanticweb.org/wiki/OWL_2

http://www.w3.org/TR/owl2-profiles/#Appendix:_Complete_Grammars_for_Profiles

http://webont.org/owled/2008/papers/owled2008eu_submission_23.pdf - Excellent paper for describing how it all works

 

Resource Description Framework

http://www.w3.org/RDF/

http://www-archive.mozilla.org/rdf/doc/examples.html - excellent example of RDF use

 

Linked Data

http://www.w3.org/TR/#tr_Linked_Data

 

First Order Logic

http://www.cs.cornell.edu/courses/cs4700/2011fa/lectures/16_FirstOrderLogic.pdf

http://mathworld.wolfram.com/First-OrderLogic.html

http://www.philosophypages.com/lg/index.htm

 

Descriptive Languages

https://en.wikipedia.org/wiki/Description_logic -Usually eschew Wikipedia for formal writing but this is a good comparison with excellent references.

 

I hope this helps you on your way…

 

Mark

 

 

From: protege-user [mailto:[hidden email]] On Behalf Of Daniel Gross
Sent: Monday, July 6, 2015 8:49 AM
To: [hidden email]
Subject: [protege-user] OWL vs. FOL

 

Hi, 

 

I have some experience with Protege 3.x, the frame based ontology building, am introducing myself (again) to OWL (and later OWL2). 

 

I am very interested in learning more about the expressiveness of OWL, OWL2 and various Description Logics (by example), as well as better understanding in what way OWL/OWL2 and DLs are less expressive than FOL.

 

I am wondering:

 

a. Is there some introductory text by example that presents the topic of expressivness of the many DL variants and in relation to FOL

b. Is there perhaps a protege "chat room" (on skype, irc, elsehwere) where I can chat about this (and other topics)

 

On a different note:

 

I also started to look at Bayesian Belief Network approach to represent causal relationships between evidences and conclusions. I also noticed works that seem to connect OWL with BBNs. How mature is such combination considered; are there any known implementations with which there exists some experience. 

 

 

thank you,

 

Daniel

 

 

--


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user




--

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: OWL vs. FOL

carter01

Daniel,

 

I agree with Pavel in his post. The situation will often drive the solution and the specific language set used. Propositional can be as expressive depending on the variables and the objective.

 

I can understand the question. First, it depends… if we use an operational ontology such as FOAF (Friend of a Friend https://svn.omdoc.org/repos/omdoc/trunk/doc/blue/foaf/note.pdf)

 

The axiom:

The foaf:Person classis a sub-class of the foaf:Agent class, since all people are considered ‘Agents’ in FOAF.

The foaf:Organization classis a sub-class of the foaf:Agent class, since all organizations are considered ‘Agents’ in FOAF.

 

This followed by:

Such that the intersection of a person and organization is a contradiction or no person is an organization and no organization is a person. A disjoint statement

 

The superclass is “Agent”, the subclasses are “Person” and “Organization”, which are disjoint meaning no instance of a person can be an organization, no instance of an organization can be a person.

 

In the FOAF Specification, an Agent is specified as:

 

Class: foaf:Agent

Agent - An agent (eg. person, group, software or physical artifact).

Status:

stable

Properties include:

gender yahooChatID account birthday icqChatID aimChatID jabberID made mbox interest tipjar skypeID topic_interest age mbox_sha1sum status msnChatID openid holdsAccount weblog

Used with:

maker member

Has Subclass

Group Person Organization

The Agent class is the class of agents; things that do stuff. A well known sub-class is Person, representing people. Other kinds of agents include Organization and Group.

The Agent class is useful in a few places in FOAF where Person would have been overly specific. For example, the IM chat ID properties such as jabberID are typically associated with people, but sometimes belong to software bots.

You should be able to reason (no pun intended) the rationale for FOL/DL/OWL usage. The logic is used in evaluation and instantiated using OWL/RDF.

 

From my experience, the implementation is not necessarily AI (in the traditional sense) but more NLP. I have formed ontologies to support text-mining, entity extraction, and decision-making using NLP processes in GATE. https://gate.ac.uk/ie/annie.html It all depends on what you are trying to do. Instantiation of an ontology framework augments the NLP to increase effectiveness. When combined with Machine Learning, the results are powerful…

 

Mark

 

From: protege-user [mailto:[hidden email]] On Behalf Of Daniel Gross
Sent: Monday, July 6, 2015 10:29 AM
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] OWL vs. FOL

 

Dear Mark,

Many thanks for the detail response.

Perhaps I should also ask about purpose and application areas for
Description Logic (DL) knowledge representations. When describing a
conceptualization of some things (say a Pizza business) using some kinds of
DL and perhaps even embodied into KBs (such as triple stores that support
DL kind of inferences). What are the typical applications areas. Is it
about standardizing definitions of terminology, or are there also
applications areas for the creation of intelligent/AI kind of systems that
take advantage of kinds of DL reasoning support?

thank you,

Daniel

p.s. I guess variants of FOL, such as implemented in Prolog as well as in
some agent-oriented logic programming platforms, have been used for more
"general purpose" intelligence/reasoning. DLs are seen as more restricted
in their application areas?

 

On Mon, Jul 6, 2015 at 6:03 PM, Mark A. Carter-Gmail <[hidden email]> wrote:

Daniel,

 

It is important to understand the differences in the terms you are using. The foundations of ontology are grounded in the metaphysical attempt to understand the nature of the “thing”. First Order Logic (FOL) is the formal mathematical system using quantified variables to define objects or the “thing”. FOL (also known as predicate logic or calculus) is distinguished from propositional logic that does not use quantified variables. Since FOL uses quantified variable, it allows for a higher degree of flexibility to describe the nature of a “thing” in essence by declaration.

 

Description Logic (DL) is a family of formal knowledge representation and as such has specific notations used for description. It is more expressive than propositional logic and more efficient than FOL due to specific operators and notations used in each individual language. The DL languages include: Attributive (AL), Frame-based (FL), and Existential (EL) that are followed by extension modifiers that define restrictions. There are several patterns that can be used in DL that are common such as, SHIQ, ALCOIN, SHOIN, SHIF, etc… Each variation of DL is formed by the annotations of the base logic followed by the restrictions used in the logic. OWL2 ontologies often expresses the “SRIQ” logic (S being the annotation for ALC or Attributive Language Complex concept negation, followed by R- limited complex role inclusion, I- Inverse properties, Q- Qualified cardinality restriction. FOL and DL can mathematically define a set of relationships but they are difficult to instantiate in a computer processing system.

 

The Web Ontology Language (OWL) is a computational markup language for expressing the complex relationships about “things” and groups of “things” within the context of the semantic web. From an IT perspective, an ontology is a specification of a conceptualization (Gruber). OWL2 provides the language for computability. A key part of developing an ontology (in OWL or any other framework) is to provide the logic that forms the relationships in a document called a specification. The logic can be annotated in FOL or DL but it must be annotated such that its explicit. There are many examples that can be used to define the process (the infamous Pizza Ontology for example). There are resource aplenty but they do take time to read, digest, and integrate into your mindscape. Here is my list, although may have some or most of these already:

 

Semantic Web

http://www.w3.org/standards/semanticweb/

http://searchsoa.techtarget.com/tip/Reviewing-the-elements-of-the-Semantic-Web

 

OWL

Note: It is important to remember that there are different grammar profiles that offer specific characteristics both normative and non-normative.

http://www.w3.org/TR/#tr_OWL_Web_Ontology_Language

http://semanticweb.org/wiki/OWL_2

http://www.w3.org/TR/owl2-profiles/#Appendix:_Complete_Grammars_for_Profiles

http://webont.org/owled/2008/papers/owled2008eu_submission_23.pdf - Excellent paper for describing how it all works

 

Resource Description Framework

http://www.w3.org/RDF/

http://www-archive.mozilla.org/rdf/doc/examples.html - excellent example of RDF use

 

Linked Data

http://www.w3.org/TR/#tr_Linked_Data

 

First Order Logic

http://www.cs.cornell.edu/courses/cs4700/2011fa/lectures/16_FirstOrderLogic.pdf

http://mathworld.wolfram.com/First-OrderLogic.html

http://www.philosophypages.com/lg/index.htm

 

Descriptive Languages

https://en.wikipedia.org/wiki/Description_logic -Usually eschew Wikipedia for formal writing but this is a good comparison with excellent references.

 

I hope this helps you on your way…

 

Mark

 

 

From: protege-user [mailto:[hidden email]] On Behalf Of Daniel Gross
Sent: Monday, July 6, 2015 8:49 AM
To:
[hidden email]
Subject: [protege-user] OWL vs. FOL

 

Hi, 

 

I have some experience with Protege 3.x, the frame based ontology building, am introducing myself (again) to OWL (and later OWL2). 

 

I am very interested in learning more about the expressiveness of OWL, OWL2 and various Description Logics (by example), as well as better understanding in what way OWL/OWL2 and DLs are less expressive than FOL.

 

I am wondering:

 

a. Is there some introductory text by example that presents the topic of expressivness of the many DL variants and in relation to FOL

b. Is there perhaps a protege "chat room" (on skype, irc, elsehwere) where I can chat about this (and other topics)

 

On a different note:

 

I also started to look at Bayesian Belief Network approach to represent causal relationships between evidences and conclusions. I also noticed works that seem to connect OWL with BBNs. How mature is such combination considered; are there any known implementations with which there exists some experience. 

 

 

thank you,

 

Daniel

 

 

--


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user

 


 

--


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: OWL vs. FOL

Pavel Klinov
In reply to this post by Daniel Gross-2
Daniel,

First, I suggest you take a look at the OWL 2 Profiles specification
[1]. Each of the profiles is a carefully designed language, based on
specific DLs, created for certain, reasonably well-defined use cases.
Those use cases can help you understand how DLs are most often used
these days.

Second, any answer to your question is debatable but here's my take.
Basically one can distinguish two major concerns:

i) Representing and reasoning about (large) schemas
ii) Dealing with data

i) is indeed mostly about standardizing terminology, as you said. Most
often it's used to represent medical or some sort of scientific
knowledge (biology, chemistry, etc,) to ensure that different
applications agree on the meaning of terms (apps like, for example,
electronic health records, medical decision support, etc.). Ontologies
like SNOMED, the Gene Ontology, etc. are widely discussed canonical
examples of this use case. The OWL 2 EL profile is the language for
this task. Classification is often the most important reasoning task.

ii) is about managing large volumes of data (Linked Data or some sort
of relational data via mappings to RDF) w.r.t. a lightweight schema.
One canonical use case is data integration, that is, managing data in
disparate data silos without physically integrating them into a single
database. A common DL ontology serves as an interface to query all
data sources via a common interface (as a virtual graph). Conjunctive
query answering, usually via SPARQL (although it's not the same
thing), is often the most important reasoning task. The OWL 2 QL and
OWL 2 RL profiles have been designed for dealing with data (they
support different features and their implementations often, but not
necessarily, are based on different techniques, i.e. query rewriting
vs. materialization).

I don't know what you mean by "intelligent AI kind of systems" but I
wouldn't say that these days DLs are used for general AI. Lots of
people would go as far as to say that logic-based AI is dead anyway. I
prefer to stay away from that debate.

HTH,
Pavel

[1] http://www.w3.org/TR/owl2-profiles/

PS. The semantic-web and public-owl-dev could be better venues for
this sort of questions. It has nothing to do with Protege per se.

On Mon, Jul 6, 2015 at 5:29 PM, Daniel Gross <[hidden email]> wrote:

> Dear Mark,
>
> Many thanks for the detail response.
>
> Perhaps I should also ask about purpose and application areas for
> Description Logic (DL) knowledge representations. When describing a
> conceptualization of some things (say a Pizza business) using some kinds of
> DL and perhaps even embodied into KBs (such as triple stores that support
> DL kind of inferences). What are the typical applications areas. Is it
> about standardizing definitions of terminology, or are there also
> applications areas for the creation of intelligent/AI kind of systems that
> take advantage of kinds of DL reasoning support?
>
> thank you,
>
> Daniel
>
> p.s. I guess variants of FOL, such as implemented in Prolog as well as in
> some agent-oriented logic programming platforms, have been used for more
> "general purpose" intelligence/reasoning. DLs are seen as more restricted
> in their application areas?
>
> On Mon, Jul 6, 2015 at 6:03 PM, Mark A. Carter-Gmail
> <[hidden email]> wrote:
>>
>> Daniel,
>>
>>
>>
>> It is important to understand the differences in the terms you are using.
>> The foundations of ontology are grounded in the metaphysical attempt to
>> understand the nature of the “thing”. First Order Logic (FOL) is the formal
>> mathematical system using quantified variables to define objects or the
>> “thing”. FOL (also known as predicate logic or calculus) is distinguished
>> from propositional logic that does not use quantified variables. Since FOL
>> uses quantified variable, it allows for a higher degree of flexibility to
>> describe the nature of a “thing” in essence by declaration.
>>
>>
>>
>> Description Logic (DL) is a family of formal knowledge representation and
>> as such has specific notations used for description. It is more expressive
>> than propositional logic and more efficient than FOL due to specific
>> operators and notations used in each individual language. The DL languages
>> include: Attributive (AL), Frame-based (FL), and Existential (EL) that are
>> followed by extension modifiers that define restrictions. There are several
>> patterns that can be used in DL that are common such as, SHIQ, ALCOIN,
>> SHOIN, SHIF, etc… Each variation of DL is formed by the annotations of the
>> base logic followed by the restrictions used in the logic. OWL2 ontologies
>> often expresses the “SRIQ” logic (S being the annotation for ALC or
>> Attributive Language Complex concept negation, followed by R- limited
>> complex role inclusion, I- Inverse properties, Q- Qualified cardinality
>> restriction. FOL and DL can mathematically define a set of relationships but
>> they are difficult to instantiate in a computer processing system.
>>
>>
>>
>> The Web Ontology Language (OWL) is a computational markup language for
>> expressing the complex relationships about “things” and groups of “things”
>> within the context of the semantic web. From an IT perspective, an ontology
>> is a specification of a conceptualization (Gruber). OWL2 provides the
>> language for computability. A key part of developing an ontology (in OWL or
>> any other framework) is to provide the logic that forms the relationships in
>> a document called a specification. The logic can be annotated in FOL or DL
>> but it must be annotated such that its explicit. There are many examples
>> that can be used to define the process (the infamous Pizza Ontology for
>> example). There are resource aplenty but they do take time to read, digest,
>> and integrate into your mindscape. Here is my list, although may have some
>> or most of these already:
>>
>>
>>
>> Semantic Web
>>
>> http://www.w3.org/standards/semanticweb/
>>
>>
>> http://searchsoa.techtarget.com/tip/Reviewing-the-elements-of-the-Semantic-Web
>>
>>
>>
>> OWL
>>
>> Note: It is important to remember that there are different grammar
>> profiles that offer specific characteristics both normative and
>> non-normative.
>>
>> http://www.w3.org/TR/#tr_OWL_Web_Ontology_Language
>>
>> http://semanticweb.org/wiki/OWL_2
>>
>>
>> http://www.w3.org/TR/owl2-profiles/#Appendix:_Complete_Grammars_for_Profiles
>>
>> http://webont.org/owled/2008/papers/owled2008eu_submission_23.pdf -
>> Excellent paper for describing how it all works
>>
>>
>>
>> Resource Description Framework
>>
>> http://www.w3.org/RDF/
>>
>> http://www-archive.mozilla.org/rdf/doc/examples.html - excellent example
>> of RDF use
>>
>>
>>
>> Linked Data
>>
>> http://www.w3.org/TR/#tr_Linked_Data
>>
>>
>>
>> First Order Logic
>>
>>
>> http://www.cs.cornell.edu/courses/cs4700/2011fa/lectures/16_FirstOrderLogic.pdf
>>
>> http://mathworld.wolfram.com/First-OrderLogic.html
>>
>> http://www.philosophypages.com/lg/index.htm
>>
>>
>>
>> Descriptive Languages
>>
>> https://en.wikipedia.org/wiki/Description_logic -Usually eschew Wikipedia
>> for formal writing but this is a good comparison with excellent references.
>>
>>
>>
>> I hope this helps you on your way…
>>
>>
>>
>> Mark
>>
>>
>>
>>
>>
>> From: protege-user [mailto:[hidden email]] On
>> Behalf Of Daniel Gross
>> Sent: Monday, July 6, 2015 8:49 AM
>> To: [hidden email]
>> Subject: [protege-user] OWL vs. FOL
>>
>>
>>
>> Hi,
>>
>>
>>
>> I have some experience with Protege 3.x, the frame based ontology
>> building, am introducing myself (again) to OWL (and later OWL2).
>>
>>
>>
>> I am very interested in learning more about the expressiveness of OWL,
>> OWL2 and various Description Logics (by example), as well as better
>> understanding in what way OWL/OWL2 and DLs are less expressive than FOL.
>>
>>
>>
>> I am wondering:
>>
>>
>>
>> a. Is there some introductory text by example that presents the topic of
>> expressivness of the many DL variants and in relation to FOL
>>
>> b. Is there perhaps a protege "chat room" (on skype, irc, elsehwere) where
>> I can chat about this (and other topics)
>>
>>
>>
>> On a different note:
>>
>>
>>
>> I also started to look at Bayesian Belief Network approach to represent
>> causal relationships between evidences and conclusions. I also noticed works
>> that seem to connect OWL with BBNs. How mature is such combination
>> considered; are there any known implementations with which there exists some
>> experience.
>>
>>
>>
>>
>>
>> thank you,
>>
>>
>>
>> Daniel
>>
>>
>>
>>
>>
>> --
>>
>> --
>>
>> Dr. Daniel Gross
>>
>> http://www.linkedin.com/in/grossd18
>>
>>
>> _______________________________________________
>> protege-user mailing list
>> [hidden email]
>> https://mailman.stanford.edu/mailman/listinfo/protege-user
>>
>
>
>
> --
>
> --
> Dr. Daniel Gross
> http://www.linkedin.com/in/grossd18
>
>
> _______________________________________________
> protege-user mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/protege-user
>
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: OWL vs. FOL

Bohms, H.M. (Michel)
Dear Daniel,

Just thank you for asking this question. The several answers provide great info and links !
Greetings Michel

Dr. ir. H.M. (Michel) Bohms
Sr. Research Scientist
Structural Reliability
T +31 (0)88 866 31 07
M +31 (0)63 038 12 20
E [hidden email]

This message may contain information that is not intended for you. If you are not the addressee or if this message was sent to you by mistake, you are requested to inform the sender and delete the message. TNO accepts no liability for the content of this e-mail, for the manner in which you use it and for damage of any kind resulting from the risks inherent to the electronic transmission of messages.

-----Original Message-----
From: protege-user [mailto:[hidden email]] On Behalf Of Pavel Klinov
Sent: woensdag 8 juli 2015 12:00
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] OWL vs. FOL

Daniel,

First, I suggest you take a look at the OWL 2 Profiles specification [1]. Each of the profiles is a carefully designed language, based on specific DLs, created for certain, reasonably well-defined use cases.
Those use cases can help you understand how DLs are most often used these days.

Second, any answer to your question is debatable but here's my take.
Basically one can distinguish two major concerns:

i) Representing and reasoning about (large) schemas
ii) Dealing with data

i) is indeed mostly about standardizing terminology, as you said. Most often it's used to represent medical or some sort of scientific knowledge (biology, chemistry, etc,) to ensure that different applications agree on the meaning of terms (apps like, for example, electronic health records, medical decision support, etc.). Ontologies like SNOMED, the Gene Ontology, etc. are widely discussed canonical examples of this use case. The OWL 2 EL profile is the language for this task. Classification is often the most important reasoning task.

ii) is about managing large volumes of data (Linked Data or some sort of relational data via mappings to RDF) w.r.t. a lightweight schema.
One canonical use case is data integration, that is, managing data in disparate data silos without physically integrating them into a single database. A common DL ontology serves as an interface to query all data sources via a common interface (as a virtual graph). Conjunctive query answering, usually via SPARQL (although it's not the same thing), is often the most important reasoning task. The OWL 2 QL and OWL 2 RL profiles have been designed for dealing with data (they support different features and their implementations often, but not necessarily, are based on different techniques, i.e. query rewriting vs. materialization).

I don't know what you mean by "intelligent AI kind of systems" but I wouldn't say that these days DLs are used for general AI. Lots of people would go as far as to say that logic-based AI is dead anyway. I prefer to stay away from that debate.

HTH,
Pavel

[1] http://www.w3.org/TR/owl2-profiles/

PS. The semantic-web and public-owl-dev could be better venues for this sort of questions. It has nothing to do with Protege per se.

On Mon, Jul 6, 2015 at 5:29 PM, Daniel Gross <[hidden email]> wrote:

> Dear Mark,
>
> Many thanks for the detail response.
>
> Perhaps I should also ask about purpose and application areas for
> Description Logic (DL) knowledge representations. When describing a
> conceptualization of some things (say a Pizza business) using some
> kinds of DL and perhaps even embodied into KBs (such as triple stores
> that support DL kind of inferences). What are the typical applications
> areas. Is it about standardizing definitions of terminology, or are
> there also applications areas for the creation of intelligent/AI kind
> of systems that take advantage of kinds of DL reasoning support?
>
> thank you,
>
> Daniel
>
> p.s. I guess variants of FOL, such as implemented in Prolog as well as
> in some agent-oriented logic programming platforms, have been used for
> more "general purpose" intelligence/reasoning. DLs are seen as more
> restricted in their application areas?
>
> On Mon, Jul 6, 2015 at 6:03 PM, Mark A. Carter-Gmail
> <[hidden email]> wrote:
>>
>> Daniel,
>>
>>
>>
>> It is important to understand the differences in the terms you are using.
>> The foundations of ontology are grounded in the metaphysical attempt
>> to understand the nature of the “thing”. First Order Logic (FOL) is
>> the formal mathematical system using quantified variables to define
>> objects or the “thing”. FOL (also known as predicate logic or
>> calculus) is distinguished from propositional logic that does not use
>> quantified variables. Since FOL uses quantified variable, it allows
>> for a higher degree of flexibility to describe the nature of a “thing” in essence by declaration.
>>
>>
>>
>> Description Logic (DL) is a family of formal knowledge representation
>> and as such has specific notations used for description. It is more
>> expressive than propositional logic and more efficient than FOL due
>> to specific operators and notations used in each individual language.
>> The DL languages
>> include: Attributive (AL), Frame-based (FL), and Existential (EL)
>> that are followed by extension modifiers that define restrictions.
>> There are several patterns that can be used in DL that are common
>> such as, SHIQ, ALCOIN, SHOIN, SHIF, etc… Each variation of DL is
>> formed by the annotations of the base logic followed by the
>> restrictions used in the logic. OWL2 ontologies often expresses the
>> “SRIQ” logic (S being the annotation for ALC or Attributive Language
>> Complex concept negation, followed by R- limited complex role
>> inclusion, I- Inverse properties, Q- Qualified cardinality
>> restriction. FOL and DL can mathematically define a set of relationships but they are difficult to instantiate in a computer processing system.
>>
>>
>>
>> The Web Ontology Language (OWL) is a computational markup language
>> for expressing the complex relationships about “things” and groups of “things”
>> within the context of the semantic web. From an IT perspective, an
>> ontology is a specification of a conceptualization (Gruber). OWL2
>> provides the language for computability. A key part of developing an
>> ontology (in OWL or any other framework) is to provide the logic that
>> forms the relationships in a document called a specification. The
>> logic can be annotated in FOL or DL but it must be annotated such
>> that its explicit. There are many examples that can be used to define
>> the process (the infamous Pizza Ontology for example). There are
>> resource aplenty but they do take time to read, digest, and integrate
>> into your mindscape. Here is my list, although may have some or most of these already:
>>
>>
>>
>> Semantic Web
>>
>> http://www.w3.org/standards/semanticweb/
>>
>>
>> http://searchsoa.techtarget.com/tip/Reviewing-the-elements-of-the-Sem
>> antic-Web
>>
>>
>>
>> OWL
>>
>> Note: It is important to remember that there are different grammar
>> profiles that offer specific characteristics both normative and
>> non-normative.
>>
>> http://www.w3.org/TR/#tr_OWL_Web_Ontology_Language
>>
>> http://semanticweb.org/wiki/OWL_2
>>
>>
>> http://www.w3.org/TR/owl2-profiles/#Appendix:_Complete_Grammars_for_P
>> rofiles
>>
>> http://webont.org/owled/2008/papers/owled2008eu_submission_23.pdf -
>> Excellent paper for describing how it all works
>>
>>
>>
>> Resource Description Framework
>>
>> http://www.w3.org/RDF/
>>
>> http://www-archive.mozilla.org/rdf/doc/examples.html - excellent
>> example of RDF use
>>
>>
>>
>> Linked Data
>>
>> http://www.w3.org/TR/#tr_Linked_Data
>>
>>
>>
>> First Order Logic
>>
>>
>> http://www.cs.cornell.edu/courses/cs4700/2011fa/lectures/16_FirstOrde
>> rLogic.pdf
>>
>> http://mathworld.wolfram.com/First-OrderLogic.html
>>
>> http://www.philosophypages.com/lg/index.htm
>>
>>
>>
>> Descriptive Languages
>>
>> https://en.wikipedia.org/wiki/Description_logic -Usually eschew
>> Wikipedia for formal writing but this is a good comparison with excellent references.
>>
>>
>>
>> I hope this helps you on your way…
>>
>>
>>
>> Mark
>>
>>
>>
>>
>>
>> From: protege-user [mailto:[hidden email]]
>> On Behalf Of Daniel Gross
>> Sent: Monday, July 6, 2015 8:49 AM
>> To: [hidden email]
>> Subject: [protege-user] OWL vs. FOL
>>
>>
>>
>> Hi,
>>
>>
>>
>> I have some experience with Protege 3.x, the frame based ontology
>> building, am introducing myself (again) to OWL (and later OWL2).
>>
>>
>>
>> I am very interested in learning more about the expressiveness of
>> OWL,
>> OWL2 and various Description Logics (by example), as well as better
>> understanding in what way OWL/OWL2 and DLs are less expressive than FOL.
>>
>>
>>
>> I am wondering:
>>
>>
>>
>> a. Is there some introductory text by example that presents the topic
>> of expressivness of the many DL variants and in relation to FOL
>>
>> b. Is there perhaps a protege "chat room" (on skype, irc, elsehwere)
>> where I can chat about this (and other topics)
>>
>>
>>
>> On a different note:
>>
>>
>>
>> I also started to look at Bayesian Belief Network approach to
>> represent causal relationships between evidences and conclusions. I
>> also noticed works that seem to connect OWL with BBNs. How mature is
>> such combination considered; are there any known implementations with
>> which there exists some experience.
>>
>>
>>
>>
>>
>> thank you,
>>
>>
>>
>> Daniel
>>
>>
>>
>>
>>
>> --
>>
>> --
>>
>> Dr. Daniel Gross
>>
>> http://www.linkedin.com/in/grossd18
>>
>>
>> _______________________________________________
>> protege-user mailing list
>> [hidden email]
>> https://mailman.stanford.edu/mailman/listinfo/protege-user
>>
>
>
>
> --
>
> --
> Dr. Daniel Gross
> http://www.linkedin.com/in/grossd18
>
>
> _______________________________________________
> protege-user mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/protege-user
>
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user