From OntologPSMW

(Difference between revisions)
Jump to: navigation, search
(Conference Call Information)
(Resources)
 
(4 intermediate revisions by one user not shown)
Line 25: Line 25:
  
 
== Agenda ==
 
== Agenda ==
* John Sowa Presentation
+
* '''John Sowa''' ''Knowledge Graphs for Language, Logic, Data, Reasoning'' [https://go.aws/2SkRNc3 Slides] [https://go.aws/2OJgblB Video Recording] [https://youtu.be/J9_kXpZAcQY YouTube Video]
 
* Discussion
 
* Discussion
  
Line 36: Line 36:
  
 
== Attendees ==
 
== Attendees ==
 +
* [[AlexShkotin|Alex Shkotin]]
 +
* [[BillHubbard|Bill Hubbard]]
 +
* [[DaveWhitten|Dave Whitten]]
 +
* [[DavidEddy|David Eddy]]
 +
* [[DougHolmes|Doug Holmes]]
 +
* [[EvanWallace|Evan Wallace]]
 +
* [[FranLightsom|Fran Lightsom]]
 +
* [[GaryBergCross|Gary Berg-Cross]]
 +
* gnk
 +
* [[HeleneFinidori|Helene Finidori]]
 +
* htd
 +
* [[JanetSinger|Janet Singer]]
 +
* [[JanetteWong|Janette Wong]]
 +
* [[JanPopiel|Jan Popiel]]
 +
* [[JessicaKent|Jessica Kent]]
 +
* [[JoanBeaudoin|Joan Beaudoin]]
 +
* [[JoelBender|Joel Bender]]
 +
* [[JohnSowa|John Sowa]]
 +
* [[KenBaclawski|Ken Baclawski]]
 +
* [[MarciaZeng|Marcia Zeng]]
 +
* [[MarcoNeumann|Marco Neumann]]
 +
* [[MarkUnderwood|Mark Underwood]]
 +
* [[MikeBennett|Mike Bennett]]
 +
* [[MinervaIntelligence|Minerva Intelligence]]
 +
* Nonesuch
 +
* [[PeteRivett|PeteRivett]]
 +
* [[RamSriram|Ram D. Sriram]]
 +
* [[RaviSharma|Ravi Sharma]]
 +
* rhmcc
 +
* [[RobSiera|Rob Siera]]
 +
* [[SeanGordon|Sean Gordon]]
 +
* [[ToddSchneider|Todd Schneider]]
 +
* [[TomTinsley|Tom Tinsley]]
  
== Proceedings ==
+
== Discussion ==
 +
[12:05] MikeBennett: http://jfsowa.com/talks/kng.pdf the link works here
 +
 
 +
[12:06] ravisharma: John I find Search or discovery missing from the title of your slides.
 +
 
 +
[12:10] David Eddy: We've lived without standards so far... why embrace them now?
 +
 
 +
[12:11] David Eddy: I like that bottom, green level... absent definitions.  Accurate reflection of reality.
 +
 
 +
[12:13] ravisharma: Slide 3 does not list target or search for KGs?
 +
 
 +
[12:17] ravisharma: Ontology is logically equivalent to Conceptual schema according to John.
 +
 
 +
[12:19] ToddSchneider: An ontology is 'realized' in some physical medium or entity. The content of an ontology is represented via some 'means', usually a language of some sort.
 +
 
 +
[12:19] David Eddy: by Tim Berners-Lee's own words... there are no semantics in his SemanticWeb... his intent (implemented today) is linked open data.
 +
 
 +
[12:21] David Eddy: "Semantic Web" was sexier marketing
 +
 
 +
[12:21] ravisharma: According to John using syntax for semantics was a mistake according to John
 +
 
 +
[12:21] DouglasRMiles: Semantic web was created due to the success of Cycorp.. we wanted to be able to use the KE tools
 +
 
 +
[12:21] DouglasRMiles: I mean have the world able to develop the types of apps Cycorp could
 +
 
 +
[12:22] DouglasRMiles: For example I before at the time could translate English into CycL
 +
 
 +
[12:22] David Eddy: By face-to-face conversation with Jim Hendler... there is no room / interest? in SemWeb stack for legacy applications.
 +
 
 +
[12:22] ravisharma: JSON is better - John
 +
 
 +
[12:23] DouglasRMiles: When the SemWeb developed RDF/OWL it was so much less expressive than CycL I just gave up
 +
 
 +
[12:23] DouglasRMiles: I mean they made most of us quit helping
 +
 
 +
[12:23] DouglasRMiles: (Developing a language that was so unexpressive)
 +
 
 +
[12:24] JoelBender: Which JSON?  Which JSON-LD?  :-/
 +
 
 +
[12:26] DouglasRMiles: I go on to say that TimBL set the world's technology 20 years back
 +
 
 +
[12:26] ravisharma: John says generate the ontology automatically rather than test it for consistency after the construction!
 +
 
 +
[12:26] DouglasRMiles: by adopting a language that may as well been SQL
 +
 
 +
[12:27] DouglasRMiles: Rather than CommonLogic or CycL
 +
 
 +
[12:27] DouglasRMiles: Well OK there was on KIF at the time
 +
 
 +
[12:27] ravisharma: CL Pat Hayes and John with others
 +
 
 +
[12:28] DouglasRMiles: *only* KIF at the time
 +
 
 +
[12:29] MikeBennett: I agree with John, I remain bemused that people doing 'ontology' in OWL assume that to be an ontology is (only) to be an OWL file, with DL, decidability etc. and anything else is 'not an ontology' - completely ignoring model theoretic aspects of the thing IMHO.
 +
 
 +
[12:32] DouglasRMiles: Yeah the whole point of having an 'ontology; is so our Rules could have a common set of Collections we referred to
 +
 
 +
[12:43] ravisharma: John asked question how to relate two representations of same concept and the transformation is Vivo Mind's future direction
 +
 
 +
[12:45] MikeBennett: Mapping remains a gnarly question in commercial deployment of ontology - although you need common meaning as the 'hub' for mapping, it's not a neat 1:1 between ontology and any database design. So the technique John is showing is important.
 +
 
 +
[12:50] ravisharma: John we want to see relationship with KG especially of NLP?
 +
 
 +
[12:52] janet singer: @Mike: How to get conceptual-level ontology work recognized and supported in its own right?
 +
 
 +
[12:54] MikeBennett: @Janet that is the existential issue I am struggling with. Corporates find it easier to treat the whole thing as an IT matter - conflating the plumbing with the fluid.
 +
 
 +
[12:54] DouglasRMiles: I bet the best way to get a ontology accepted is to ship a deliverable that can use it
 +
 
 +
[12:55] MikeBennett: Organizations need to understand that developing conceptual ontology is a distinct skill-set, just like a lawyer or accountant.
 +
 
 +
[12:56] MikeBennett: @Douglas absolutely. Need to refine methodologies for extracting operational ontologies and / or data models from conceptual ontology, per context (where the context is the use case of the application). Left to IT, there are naive assumptions that one kind of ontology can do it all.
 +
 
 +
[12:57] Mark Underwood @knowlengr: @Janet @Mike Isn't it all solved as Aspect-oriented ?
 +
 
 +
[12:58] MikeBennett: Hmmm
 +
 
 +
[12:58] David Eddy: Is software (from source code to executables) considered to be "documents?"
 +
 
 +
[12:59] DouglasRMiles: John agrees with that me that ontology  was to make existing powerful system scale
 +
 
 +
[13:00] MikeBennett: Aspect Oriented seems a good name for what I was thinking of. Needs to be fleshed out for ontology. Then we can deliver usable applications and demonstrate the relevance of conceptual ontology (actually basic model driven development, in a different set of syntaxes).
 +
 
 +
[13:00] DouglasRMiles: existing powerful systems = the AI software being creating by contractors
 +
 
 +
[13:01] janet singer: I think the relationships between concepts, representations and reality have gotten more muddled since the 1970s as people have tried to be more ontologically profound. Would it help to go back to the earlier conceptual schemas language?
 +
 
 +
[13:01] ravisharma: @Todd - thanks, great question
 +
 
 +
[13:03] ToddSchneider: Sorry,late for another Zoom meeting; Have to go. Thank you.
 +
 
 +
[13:04] Helene Finidori: To add to Janet's point, isn't it a major issue to focus on syntax / structure, more than on meaning? This was mentioned by John at the beginning of the presentation: the issue of syntax as a basis of semantics...
 +
 
 +
[13:04] ravisharma: john said KGs are doing informal and approx computation
 +
 
 +
[13:05] David Eddy: to be continued...
 +
 
 +
[13:06] MikeBennett: @Helene exactly. As I say to clients: 'Repeat after me: Syntax is not semantics!'
 +
 
 +
[13:06] David Eddy: @MikeB... "syntax is not semantics"  Excellent
 +
 
 +
[13:07] AlexShkotin: David Eddy, you should raise hand here, pls.
 +
 
 +
[13:07] ravisharma: primary info is derived from semantics, and follow on Q is imp
 +
 
 +
[13:07] janet singer: @Mark From WP: The most basic criticism of the effect of AOP is that control flow is obscured, and that it is not only worse than the much-maligned GOTO, but is in fact closely analogous to the joke COME FROM statement[11].
 +
 
 +
[13:09] Joan Beaudoin: Thank you!
 +
 
 +
[13:09] JoelBender: Thank you!
 +
 
 +
[13:09] Helene Finidori: Thank you! Could someone put the URL of John's slides here? Thanks!
 +
 
 +
[13:10] janet singer: http://jfsowa.com/talks/kng.pdf
 +
 
 +
[13:11] KenBaclawski: John may be modifying the slides, so to see the original you can find it at https://go.aws/2SkRNc3
 +
 
 +
[13:11] janet singer: Analogies: a topic to be covered later in the series
 +
 
 +
[13:15] ravisharma: Janet thanks, I tried asking semantics as a starting point for the stack!
  
 
== Resources ==
 
== Resources ==
 +
* [https://go.aws/2SkRNc3 Slides]
 +
* [https://go.aws/2OJgblB Video Recording]
 +
* [https://youtu.be/J9_kXpZAcQY YouTube Video]
  
 
== Previous Meetings ==
 
== Previous Meetings ==

Latest revision as of 14:48, 12 February 2020

[ ]
    (1)
Session John Sowa
Duration 1 hour
Date/Time 12 February 2020 17:00 GMT
9:00am PST/12:00pm EST
5:00pm GMT/6:00pm CET
Convener KenBaclawski
Track Whence

Contents

Knowledge graphs, closely related to ontologies and semantic networks, have emerged in the last few years to be an important semantic technology and research area. As structured representations of semantic knowledge that are stored in a graph, KGs are lightweight versions of semantic networks that scale to massive datasets such as the entire World Wide Web. Industry has devoted a great deal of effort to the development of knowledge graphs, and they are now critical to the functions of intelligent virtual assistants such as Siri and Alexa. Some of the research communities where KGs are relevant are Ontologies, Big Data, Linked Data, Open Knowledge Network, Artificial Intelligence, Deep Learning, and many others.     (2A)

[edit] Conference Call Information     (2C)

[edit] Attendees     (2D)

[edit] Discussion     (2E)

[12:05] MikeBennett: http://jfsowa.com/talks/kng.pdf the link works here     (2E1)

[12:06] ravisharma: John I find Search or discovery missing from the title of your slides.     (2E2)

[12:10] David Eddy: We've lived without standards so far... why embrace them now?     (2E3)

[12:11] David Eddy: I like that bottom, green level... absent definitions. Accurate reflection of reality.     (2E4)

[12:13] ravisharma: Slide 3 does not list target or search for KGs?     (2E5)

[12:17] ravisharma: Ontology is logically equivalent to Conceptual schema according to John.     (2E6)

[12:19] ToddSchneider: An ontology is 'realized' in some physical medium or entity. The content of an ontology is represented via some 'means', usually a language of some sort.     (2E7)

[12:19] David Eddy: by Tim Berners-Lee's own words... there are no semantics in his SemanticWeb... his intent (implemented today) is linked open data.     (2E8)

[12:21] David Eddy: "Semantic Web" was sexier marketing     (2E9)

[12:21] ravisharma: According to John using syntax for semantics was a mistake according to John     (2E10)

[12:21] DouglasRMiles: Semantic web was created due to the success of Cycorp.. we wanted to be able to use the KE tools     (2E11)

[12:21] DouglasRMiles: I mean have the world able to develop the types of apps Cycorp could     (2E12)

[12:22] DouglasRMiles: For example I before at the time could translate English into CycL     (2E13)

[12:22] David Eddy: By face-to-face conversation with Jim Hendler... there is no room / interest? in SemWeb stack for legacy applications.     (2E14)

[12:22] ravisharma: JSON is better - John     (2E15)

[12:23] DouglasRMiles: When the SemWeb developed RDF/OWL it was so much less expressive than CycL I just gave up     (2E16)

[12:23] DouglasRMiles: I mean they made most of us quit helping     (2E17)

[12:23] DouglasRMiles: (Developing a language that was so unexpressive)     (2E18)

[12:24] JoelBender: Which JSON? Which JSON-LD?  :-/     (2E19)

[12:26] DouglasRMiles: I go on to say that TimBL set the world's technology 20 years back     (2E20)

[12:26] ravisharma: John says generate the ontology automatically rather than test it for consistency after the construction!     (2E21)

[12:26] DouglasRMiles: by adopting a language that may as well been SQL     (2E22)

[12:27] DouglasRMiles: Rather than CommonLogic or CycL     (2E23)

[12:27] DouglasRMiles: Well OK there was on KIF at the time     (2E24)

[12:27] ravisharma: CL Pat Hayes and John with others     (2E25)

[12:28] DouglasRMiles: *only* KIF at the time     (2E26)

[12:29] MikeBennett: I agree with John, I remain bemused that people doing 'ontology' in OWL assume that to be an ontology is (only) to be an OWL file, with DL, decidability etc. and anything else is 'not an ontology' - completely ignoring model theoretic aspects of the thing IMHO.     (2E27)

[12:32] DouglasRMiles: Yeah the whole point of having an 'ontology; is so our Rules could have a common set of Collections we referred to     (2E28)

[12:43] ravisharma: John asked question how to relate two representations of same concept and the transformation is Vivo Mind's future direction     (2E29)

[12:45] MikeBennett: Mapping remains a gnarly question in commercial deployment of ontology - although you need common meaning as the 'hub' for mapping, it's not a neat 1:1 between ontology and any database design. So the technique John is showing is important.     (2E30)

[12:50] ravisharma: John we want to see relationship with KG especially of NLP?     (2E31)

[12:52] janet singer: @Mike: How to get conceptual-level ontology work recognized and supported in its own right?     (2E32)

[12:54] MikeBennett: @Janet that is the existential issue I am struggling with. Corporates find it easier to treat the whole thing as an IT matter - conflating the plumbing with the fluid.     (2E33)

[12:54] DouglasRMiles: I bet the best way to get a ontology accepted is to ship a deliverable that can use it     (2E34)

[12:55] MikeBennett: Organizations need to understand that developing conceptual ontology is a distinct skill-set, just like a lawyer or accountant.     (2E35)

[12:56] MikeBennett: @Douglas absolutely. Need to refine methodologies for extracting operational ontologies and / or data models from conceptual ontology, per context (where the context is the use case of the application). Left to IT, there are naive assumptions that one kind of ontology can do it all.     (2E36)

[12:57] Mark Underwood @knowlengr: @Janet @Mike Isn't it all solved as Aspect-oriented ?     (2E37)

[12:58] MikeBennett: Hmmm     (2E38)

[12:58] David Eddy: Is software (from source code to executables) considered to be "documents?"     (2E39)

[12:59] DouglasRMiles: John agrees with that me that ontology was to make existing powerful system scale     (2E40)

[13:00] MikeBennett: Aspect Oriented seems a good name for what I was thinking of. Needs to be fleshed out for ontology. Then we can deliver usable applications and demonstrate the relevance of conceptual ontology (actually basic model driven development, in a different set of syntaxes).     (2E41)

[13:00] DouglasRMiles: existing powerful systems = the AI software being creating by contractors     (2E42)

[13:01] janet singer: I think the relationships between concepts, representations and reality have gotten more muddled since the 1970s as people have tried to be more ontologically profound. Would it help to go back to the earlier conceptual schemas language?     (2E43)

[13:01] ravisharma: @Todd - thanks, great question     (2E44)

[13:03] ToddSchneider: Sorry,late for another Zoom meeting; Have to go. Thank you.     (2E45)

[13:04] Helene Finidori: To add to Janet's point, isn't it a major issue to focus on syntax / structure, more than on meaning? This was mentioned by John at the beginning of the presentation: the issue of syntax as a basis of semantics...     (2E46)

[13:04] ravisharma: john said KGs are doing informal and approx computation     (2E47)

[13:05] David Eddy: to be continued...     (2E48)

[13:06] MikeBennett: @Helene exactly. As I say to clients: 'Repeat after me: Syntax is not semantics!'     (2E49)

[13:06] David Eddy: @MikeB... "syntax is not semantics" Excellent     (2E50)

[13:07] AlexShkotin: David Eddy, you should raise hand here, pls.     (2E51)

[13:07] ravisharma: primary info is derived from semantics, and follow on Q is imp     (2E52)

[13:07] janet singer: @Mark From WP: The most basic criticism of the effect of AOP is that control flow is obscured, and that it is not only worse than the much-maligned GOTO, but is in fact closely analogous to the joke COME FROM statement[11].     (2E53)

[13:09] Joan Beaudoin: Thank you!     (2E54)

[13:09] JoelBender: Thank you!     (2E55)

[13:09] Helene Finidori: Thank you! Could someone put the URL of John's slides here? Thanks!     (2E56)

[13:11] KenBaclawski: John may be modifying the slides, so to see the original you can find it at https://go.aws/2SkRNc3     (2E58)

[13:11] janet singer: Analogies: a topic to be covered later in the series     (2E59)

[13:15] ravisharma: Janet thanks, I tried asking semantics as a starting point for the stack!     (2E60)

[edit] Resources     (2F)

[edit] Previous Meetings     (2G)


[edit] Next Meetings     (2H)