From OntologPSMW

Jump to: navigation, search
[ ]
    (1)
Session Synthesis III
Duration 1 hour
Date/Time 10 June 2020 16:00 GMT
9:00am PDT/12:00pm EDT
5:00pm BST/6:00pm CEST
Convener KenBaclawski

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)

Agenda     (2B)

Conference Call Information     (2C)

Attendees     (2D)

Proceedings     (2E)

[12:03] Ravi Sharma: Here is an outline to consider:     (2E1)

[12:13] Mike Bennett: Not sure why standards come under the 'whither' heading. Wouldn't that be about future directions (which would include the question of when and whether to standardize)     (2E3)

[12:15] David Eddy: My question would be... how (or can) KGs be used to support legacy software portfolios?     (2E4)

[12:15] Todd Schneider: So, 'DIFFERENT VIEWS OF KG' is the whence introduction?     (2E5)

[12:16] Mark Underwood @knowlengr: Have no solution, but many presenters drifted quite far from knowledge graphs, which has a specific meaning to industry. That's OK perhaps, but we should gain consensus on remaining more inward- than outward-looking with the communique     (2E6)

[12:16] Mark Underwood @knowlengr: The "drift" often tended to be topics from previous years     (2E7)

[12:17] Todd Schneider: Mark, I thought one 'item' we saw was that there was no single 'understanding' of what a knowledge graph maybe.     (2E8)

[12:18] Todd Schneider: After an introduction (i.e. the 'Whence'), a section that describes current uses (i.e. what and maybe parts of why).     (2E9)

[12:20] Todd Schneider: Whither: what is the likely future of     (2E10)

[12:21] Ravi Sharma: David - thanks I did     (2E11)

[12:24] Mike Bennett: Have we covered 'Research challenges' and 'Future Directions'?     (2E12)

[12:25] Mike Bennett: There was a definition of KGs proposed at the KG Conference     (2E13)

[12:26] Mike Bennett: 'Standards' would be a good heading for standards     (2E14)

[12:28] Ravi Sharma: David - I think John Sowa and Mike and Mark did and regularly touch on mainframe cobol and other legacy integration questions.     (2E15)

[12:29] Ravi Sharma: But my Q is whether we discussed this or is it an outward view?     (2E16)

[12:29] Mike Bennett: I never mentioned Cobol.     (2E17)

[12:31] Mark Underwood @knowlengr: David: Agree about Main Frame COBOL, though I am an ex VAX COBOL developer. Legacy apps are being addressed in some KG efforts but it's not well covered. And I don't think any presenters advanced the cause, as you note     (2E18)

[12:32] Ravi Sharma: Since i assume that there are no major standards, These are futuristic or in-works standards like financial or search that KGs standards are relevant to but some people only described their or standard body's work but did not always integrate them to KGs?     (2E19)

[12:32] Mark Underwood @knowlengr: Some legacy apps are impacted by Mike B's presentation on FIBO though     (2E20)

[12:32] Todd Schneider: Sections:     (2E21)

  • Introduction     (2E22)
  • Whence - what brought the use of graphs in as persistence mechanisms (will need to address other non-relational persistence mechanism); historical background; this can include parts of 'why'     (2E23)
  • Current State - How are graph persistence mechanisms being used     (2E24)
  • Problems - what are the differences in the current uses and what problems they may cause going forward     (2E25)
  • Whither - Our recommendation of how a the notion 'knowledge graph' should be defined; How standards can help (and maybe a lead in to the next year's summit topic).     (2E26)

[12:33] David Eddy: @Mark... particularly from Mike B's side with emphasis on FIBO & financial services... HUGH chaos in dozens & dozens of languages & DBMSs in use.     (2E27)

[12:33] Mike Bennett: Proposal: assuming some of our presentations did touch on what is actually being done - then we should have a section on how KGs are being put to work in a practical setting.     (2E28)

[12:33] Gary Berg-Cross: All 3 of the presenters at June 2nd, 2020 made points on What are some open research questions on knowledge graphs: Richard Socher, Mark Musen, RV Guha     (2E29)

[12:34] Ravi Sharma: No Mike I meant that you John and Mike describe how we integrate KGs concepts or ontologies and address Ontologies including cobol and mainframe by John Sowa. I should name each speaker and subject but mostly legacy integration?     (2E30)

[12:34] Gary Berg-Cross: These presentations I mentioned are from the Stanford course - Knowledge Graphs How should AI explicitly represent knowledge?     (2E31)

[12:35] David Eddy: @Ravi... INTEGRATION is a dirty word in the game... INTEROPERABLE is the word. The silos will NEVER go away... they need to learn to take with each other.     (2E32)

[12:35] Mike Bennett: Can we make cross-reference to Stanford CS520 and / or the KG Conf in our Communique?     (2E33)

[12:36] Mark Underwood @knowlengr: @Mike +1 A sort of "Current applications" section perhaps     (2E34)

[12:36] David Eddy: @Mike... are the CS520 sessions online? I only tuned in to the very last session.     (2E35)

[12:36] Mike Bennett: A lot of what we label as ontology is more accurately epistemology. Particularly KGs, where there is instance data (what we know) overlaid on what it means.     (2E36)

[12:37] Ravi Sharma: Ken - good point what is Knowledge whether OKN or not. At least domain knowledge, knowledge bases and modern Knowledge management systems should have been covered in details before describing their Graphical views or apps?     (2E37)

[12:38] Gary Berg-Cross: @ Mike, yes I plan on citing the Stanford course - Knowledge Graphs. For example What is a knowledge graph?Denny Vrandei Jans Aasman Mikhail Galkin     (2E38)

[12:38] Alex Shkotin: @David youtube cs520.     (2E39)

[12:39] Ravi Sharma: Janet said - use some of them (speakers contents) in filling gaps.about knowledge - history to present!     (2E40)

[12:39] Gary Berg-Cross: Mark Musen leverage Newell's idea of The Knowledge level.     (2E41)

[12:39] Mike Bennett: @David I believe the materials were only formally made available to registered students. Some people have been clipping screenshots but we would not be able to cite those. We can approach specific presenters for their slides maybe     (2E42)

[12:39] David Eddy: @Todd... do remember, Tim Berners-Lee stating that "semantic web" was a sexy marketing label for "linked data"     (2E43)

[12:39] Bobbin Teegarden: Knowledge Graph Conference: We are excited to share with you our KGC 2020 media library which contains all the recorded content from the event. We encourage you to access and browse the content. Please access the link here: https://conference.knowledgegraph.tech/Profile/Verification?id=94s4HZ5ZyBOmcTV2PFW6AbTL7I3vXKQlggwpxu9yGwE%3d     (2E44)

[12:41] Ravi Sharma: Janet said - look at KGs and ontologies in both as well as complementarity.     (2E45)

[12:42] Alex Shkotin: @Bobbin your URL asks to create your account - is that OK?     (2E46)

[12:42] Todd Schneider: David, no. I'm not familiar with that assertion.     (2E47)

[12:43] Todd Schneider: Ravi, we can 'say more' than what was presented during this summit.     (2E48)

[12:43] Alex Shkotin: @Mike & @David, what look here https://www.youtube.com/channel/UC0NGAUV_9DASb0gQE3Ov9ZA     (2E49)

[12:44] David Eddy: John Sowa's oft repeated point... SemanticWeb, DARPA stack (triples, RDF, etc.) is simply yet another silo, unless it helps with the undocumented legacy systems that enable society.     (2E50)

[12:45] Janet Singer: @Mike I think CS520 changed their policy so the slides are available here https://web.stanford.edu/class/cs520/     (2E51)

[12:45] John Sowa: Last week, I gave a talk at the European Semantic Web Conf. (virtualized)     (2E52)

[12:46] John Sowa: I extended the slides with more material that addresses many of the questions on this list.     (2E53)

[12:46] Janet Singer: Here are Mark Musen's slides https://web.stanford.edu/class/cs520/abstracts/musen.html     (2E55)

[12:47] Ravi Sharma: Ravi agreed to use his chat suggested outline and will also use thoughts presented in today's chat to come up with a set of slides that will perhaps help us organize the topics or set of slides next week. sections for Communique. Will start first.     (2E56)

[12:47] Gary Berg-Cross: It is interesting that the world of Network Science, especially the work on dynamic systems in Europe, is not a consideration as related to the potential for evaluating KG from the perspective of useful metrics.     (2E57)

[12:49] Ravi Sharma: Gary will submit Stanford work (Class) etc. and submit in next two weeks.     (2E58)

[12:49] David Eddy: @Ken... a major issue here is academic research is constrained to citable research. The legacy systems are proprietary "secrets" behind the corporate firewalls... very difficult to get access to.     (2E59)

[12:49] Ravi Sharma: these are notes.     (2E60)

[12:51] Ravi Sharma: David-do you call all working applications as legacy vs futuristic upgrades and developments of those?     (2E62)

[12:52] David Eddy: @Ravi... a collection of working legacy applications is a constantly evolving work... so distinctions of "future" is simply part of what happens.     (2E63)

[12:54] David Eddy: @Mark... what about the CONTENTS that might be presented in KGs? Where does the content come from?     (2E64)

[12:54] Todd Schneider: We'll need say a bit about the use of reasoners.     (2E65)

[12:56] Gary Berg-Cross: The May 26th, 2020 session of the Stanford course had examples -What are some high value use cases of knowledge graphs? Jay Yu Apoorv Saxena David Newman     (2E66)

[12:56] Todd Schneider: Instead of Neo4J try Parliament: https://github.com/SemWebCentral/parliament/releases     (2E67)

[12:58] Mark Underwood @knowlengr: All Thx for the suggestions     (2E68)

[12:58] John Sowa: Graphs and logic are semantically *identical*. Inside the computer, there is no difference in the options for representing either or both.     (2E69)

[12:59] Mike Bennett: It turns out real people quite like the idea of graphs for visualization of how various things are connected in the world.     (2E70)

[13:00] David Eddy: two standards book... "Philadelphia's Philosopher Mechanics: A History of the Franklin Institute 1824 - 1865" by Bruce Sinclair     (2E71)

[13:00] Ravi Sharma: David - is every working App a legacy or do you have a definition and some examples?     (2E72)

[13:01] Gary Berg-Cross: I took an action item to try to abstract relevant material from the Stanford course on KGs as input to the Communique.     (2E73)

[13:02] David Eddy: Open Standards & the Digital Age: History, Ideology, & Networks. By Andrew L. Russell     (2E74)

[13:03] Gary Berg-Cross: I also suggested that of us that recruited a speaker briefly summarize the main points of the speaker....and we should also leverage the key points in chat from prior synthesis sessions.     (2E75)

[13:03] Mark Underwood @knowlengr: @Gary is the overlapping with what I just agreed to cover? Search for applied, current engineering projects?     (2E76)

[13:03] David Eddy: @Ravi... yes... a working application is legacy... because the person/team/sponsor will move to other activities & expect the "legacy" to do something useful     (2E77)

[13:03] Mark Underwood @knowlengr: @Mike Viz for CL problematic     (2E78)

[13:03] Gary Berg-Cross: @Mark...it depends on what sources you might use.     (2E79)

[13:03] Janet Singer: @Gary, @Mark: Maybe you both works from the CS520 material, with Mark on applications and Gary on broader issues     (2E80)

[13:04] Mark Underwood @knowlengr: @Janet @Gary Sure     (2E81)

[13:04] Janet Singer: There's enough good material for both     (2E82)

[13:05] David Eddy: "Engineering Rules" by Yates & Murphy     (2E83)

[13:05] Mark Underwood @knowlengr: Meeting conflict must exit. Be well, all.     (2E84)

[13:06] David Eddy: ...yes... to be continued...     (2E85)

[13:09] Ravi Sharma: David - thanks for clarity, How do we handle current developments that are often in parallel and it's like hitting a flying object while moving! and so interoperation and integration of software is desirable but often less affordable?     (2E86)

[13:10] Ravi Sharma: Ken Thanks,     (2E87)

Resources     (2F)

Previous Meetings     (2G)


Next Meetings     (2H)