Session	Introductory
Duration	1 hour
Date/Time	October 30 2019 16:00 GMT
	9:00am PDT/12:00pm EDT
	4:00pm GMT/5:00pm CET
Convener	Ken Baclawski

Ontology Summit 2020 What is new about knowledge graphs?

Agenda

This is a followup session to continue the discussion from 4 September 2019.
Please note that the US and Canada are still observing Daylight Saving Time, while Europe is not.
Video Recording
YouTube Video

Conference Call Information

Date: Wednesday, 30-October-2019
Start Time: 9:00am PDT / 12:00pm EDT / 5:00pm CET / 4:00pm GMT / 1600 UTC
- ref: World Clock
Expected Call Duration: 1 hour

The Video Conference URL is https://zoom.us/j/689971575
- iPhone one-tap :
  - US: +16699006833,,689971575# or +16465588665,,689971575#
- Telephone:
  - Dial(for higher quality, dial a number based on your current location): US: +1 669 900 6833 or +1 646 558 8665
  - Meeting ID: 689 971 575
  - International numbers available: https://zoom.us/u/Iuuiouo
Chat Room

Participants

Proceedings

[12:12] David Eddy: Do let us remember, TBL's intent was "connected data" which was deemed unacceptable... Semantic Web was sexy. Voila!

[12:13] David Eddy: "unacceptable" meaning not sexy enough

[12:14] David Eddy: @Janet - thumbs up

[12:15] David Eddy: @Paul.... yes poorly documented, deeply buried "ontologies"

[12:15] David Eddy: @Paul... but "agreement" by organizational silos

[12:16] David Eddy: where everyone brings their own understandings & definitions

[12:21] janet singer: Paul, Todd, Ken are discussing issues of constraints on generalization

[12:21] John Sowa: I very strongly agree that ontologies are a BARRIER to interoperability among independently developed systems.

[12:21] David Eddy: U-G-L-Y!!!!

[12:22] David Eddy: code & data in the wild... I can't find what's there, so I just make up a new one, & a new one & an new one...

[12:23] David Eddy: How about picking on Pl/1 rather than COBOL?

[12:24] David Eddy: @Todd... but who's going to PAY for the documentation & keep it updated?

[12:26] Mark Underwood: Fashionable or not, the knowledge graph approaches are not only RDF/OWL/Semantic web based, as noted in this piece which was one of the reasons we chose this as a topic http://bit.ly/2LkAbKj

[12:26] Mark Underwood: no, not that URL

[12:27] David Eddy: Let's remember the scale we're dealing with... for a small Fortune 350 company, with 10,000 employees... entry point with just IBM mainframes, there are likely 1,000+ applications.

[12:28] Mark Underwood: https://www.freecodecamp.org/news/user-interface-test-automation-3da36b132077/

[12:29] Mark Underwood: This is the ACM article https://queue.acm.org/detail.cfm?id=3332266

[12:33] Mark Underwood: FIBO API's are a point of commonmality in finance automation, albeit still lightly adopted, has traction https://spec.edmcouncil.org/fibo/ especially in mortgage lending and some parts of securities systems

[12:34] janet singer: Ontological thinking as opposed to ontologies

[12:34] John Sowa: I strongly support standards. They become islands of shared ontology.

[12:37] janet singer: Ontological thinking is what Paul was doing by recognizing implicit ontologies to develop data-driven bridges

[12:38] Paul Tyson: There's no requirement for a knowledge graph (aka RDF dataset) to follow an explicit ontology. The useful use cases for KGs go far beyond ontologies.

[12:39] janet singer: @Todd - those methods/questions/distinctions would be for ontological thinking, not ontologies as fixed products

[12:39] ToddSchneider: Janet, why not?

[12:40] David Eddy: but so called enterprise architects IGNORE the AS IS decades of poorly documented operational systems

[12:40] ToddSchneider: David, I was referring to an architecture for ontologies.

[12:42] ToddSchneider: Janet, the scope (of an analysis) and the actual analysis are two different aspects of ontology development.

[12:43] David Eddy: Ok... instead of dumping on COBOL, let's pick on TPF

[12:44] ToddSchneider: TPF?

[12:44] Mark Underwood: LinkedIn's non-RDF knowledge graph https://engineering.linkedin.com/blog/2016/10/building-the-linkedin-knowledge-graph

[12:44] David Eddy: TPF is the Assembler like language SABRE is written in.

[12:44] Douglas R. Miles: Here was a KG system ( https://subversion.assembla.com/svn/cogchar/trunk/src_resources_core/org/cogchar/onto/ ) I created for robotic configurations http://www.cogchar.org/

[12:45] Douglas R. Miles: How it was developed is the application was completely functional before I moved the config/data to RDF

[12:46] David Eddy: What's "bad" is the implicit assumption there is ONE, UNIVERSAL...

[12:46] Douglas R. Miles: Ideally disparate applications Cogchar and AppDatper were able to interoperate

[12:47] Mark Underwood: Douglas "PushyScreen" ? <g> https://subversion.assembla.com/svn/cogchar/trunk/src_resources_core/org/cogchar/onto/PushyScreen_owl2.n3

[12:49] Douglas R. Miles: PushyScreen configured what UIs would appear on the robot's Chest Area

[12:49] Ken Baclawski: Janet Singer said: "The data is precise, but the interpretation is vague."

[12:49] Douglas R. Miles: So if a user wanted to add a new feature .. they could assert the feature into the SPARQL database

[12:50] Ken Baclawski: Several participants brought up the idea of use cases. Could be try to collect use cases?

[12:50] Douglas R. Miles: Sorted also in the Graphs was animations the robot could carry out

[12:51] janet singer: Paul: As a community we should identify the use cases for knowledge graphs

[12:51] BobbinTeegarden: @John, I couldn't agree more about the desirability of vagueness to make knowledge reuseable and ready to apply, it's the same with business process models: the deeper you go the less reuseable they become, and the more brittle.

[12:51] Mark Underwood: I have previously offered a family of use cases for cybersec mappings to the DMTF framework for IT infrastructure

[12:51] David Eddy: I most strongly object to the use of "software engineering" Software is more like the creative process of making movies. Are there such things (titles) as "Movie Engineer?"

[12:52] Mark Underwood: A la microtheories, the definitional breadth is offered by DMTF, which isn't an ontology ... There are a few cybersec ontologies but they are very fragile because they aren't mapped to an IT framework like DMTF, which defines for instance what a power supply in a server is.

[12:53] RaviSharma: John - My Q is how the information in Dictionary is used for ontology?

[12:53] Ken Baclawski: @David Eddy: There are "Media Engineers".

[12:53] Mark Underwood: https://www.dmtf.org/standards/redfish

[12:54] David Eddy: "Media Engineer"... first time I've seen that title

[12:55] BobbinTeegarden: @John: Implied ontology, like Boehm's implicate order... ?

[12:56] Ken Baclawski: @David Eddy: Try a search for "media engineer salary".

[12:56] Paul Tyson: If everything is an ontology, what's the point of using the word? Seems to me the essence of ontology is type and attribute definitions and relationships. A KG that includes neither doesn't qualify as an ontology.

[12:56] David Eddy: Software engineer = 117,000,000. Media engineer = 72,000 Per Dr Google

[12:57] BobbinTeegarden: Implicate order and explicate order are ontological concepts for quantum theory coined by theoretical physicist David Bohm during the early 1980s. They are used to describe two different frameworks for understanding the same phenomenon or aspect of reality.Wikipedia

[12:57] janet singer: The explicit vs implicit ontology distinction is very useful

[12:58] janet singer: But it is not true that each individual has a single fixed ontology as opposed to an ecosystem of loosely compatible ontological assumptions and the ability to negotiate and evolve as needed

[12:58] ToddSchneider: Have to go. Thank you.

[13:00] Mark Underwood: Also must go ... Let me know if we choose to flesh out some use cases DMTF may be too gnarly and non-ontological but is the most rich

[13:01] janet singer: Use cases would definitely be a good target for the communique

[13:07] janet singer: John says: Ontologists should develop tools to support people doing the ontological thinking they need to solve specific real problems

[13:08] David Eddy: @Janet: are there tools that automagically extract Ontology from working code?

[13:10] janet singer: @David: John could probably address that wrt Vivomind

[13:12] Paul Tyson: John heard Doug(?) describe a robotics RDF application and took away from it Ontology. I heard the same description and took away creative use of RDF datasets, by reading certain subgraphs of interest.

[13:12] RaviSharma: Thanks to all hoping to participate from India next week

[13:13] janet singer: @David: My own view is that the challenge is not extracting one or more ontological theories from code, data, observed human behavior, etc., but how useful whatever you abstract would be

[13:13] John Sowa: David and Janet, look at the examples in http://jfsowa.com/talks/cogmem.pdf

[13:14] David Eddy: @JFS... thx

[13:15] David Eddy: ...or to be a serf

[13:15] janet singer: @David: What would the scope of application or constraints on generalization of that theory be

[13:16] David Eddy: @Janet... to be continued.

Resources

Previous Meetings

... further results

Next Meetings