Actions

Ontolog Forum

Session Topic Discussion
Duration 1 hour
Date/Time August 12 2020 16:00 GMT
9:00am PDT/12:00pm EDT
5:00pm BST/6:00pm CEST
Convener Ken Baclawski

Ontology Summit 2021 Topic Discussion

Agenda

Open-ended discussion of the topics for the next summit.

Some possible topics:

Video Recording

Conference Call Information

  • Date: Wednesday, 12-August-2020
  • Start Time: 9:00am PDT / 12:00pm EDT / 6:00pm CEST / 5:00pm BST / 1600 UTC
  • Expected Call Duration: 1 hour
  • The Video Conference URL is https://bit.ly/3i1uPRl
    • iPhone one-tap :
      • +16465588656,,83077436914#,,,,,,0#,,822275# US (New York)
      • +13017158592,,83077436914#,,,,,,0#,,822275# US (Germantown)
    • Telephone:
  • Chat Room: https://bit.ly/39PzQJW
    • If the chat room is not available, then use the Zoom chat room.

Participants

Discussion

[12:22] MikeBennett: This harmonizing / gradient idea feeds nicely into our SWAO idea of trying to better characterize ontologies. Eight now a lot of people are using terminologies instead of concept ontology because they've been told ontology is an OWL file for an application for inference processing, only.

[12:24] Russell Reinsch: Data doesn't operate. Cannot interoperate.

[12:25] Russell Reinsch: Is harmonizing more like normalizing?

[12:28] Douglas R. Miles: My mic isn't working so I will need to use chat to talk

[12:29] Douglas R. Miles: But I feel these topics are specific enough they are definitely ideal for presentations.. but a Theme they are very specific

[12:29] Ken Baclawski: @[12:24] Russell Reinsch: I was thinking in terms of "Ontology Summit 2016 : Framing the Conversation: Ontologies within Semantic Interoperability Ecosystems"

[12:30] Ken Baclawski: Possible topic: "Harmonizing Semantic Resources"

[12:30] Russell Reinsch: Ok

[12:31] Douglas R. Miles: I agree with what Russel said, as a General theme we could reuse "Harmonizing Semantic Resources"

[12:33] Douglas R. Miles: (I am grateful still of Gary is the one putting forth ideas we can think about)

[12:35] MikeBennett: Ontology Engineering is the key term here. Our remit is to show some leadership in recognizing and articulating new things that contribute to that discipline.

[12:36] MikeBennett: The interoperability agenda is something we have already covered. We're not here to teach what is already known but to research.

[12:39] Douglas R. Miles: "Tools that Ontologist may use to work together"

[12:40] MikeBennett: Address why domain people are still building controlled vocabularies? What don't they know about concept formation?

[12:40] MikeBennett: +1 to everything on Gary's 2nd slide

[12:43] DrRaviSharma: Ken please show what Gary sent rather than peoples images on zoom

[12:45] Ram D.Sriram: Learning Ontologies could be a topic

[12:47] BobbinTeegarden: Could interoperability please include human interfaces to ontologies: v

[12:48] MikeBennett: @Bobbin +1

[12:49] BobbinTeegarden: Human simple browsing and visualization (of the full graph), and query with something SIRI-like (simple, not SQLlike)

[12:51] Ram D.Sriram: Another topic: Neurosymbolic Computing

[12:51] BobbinTeegarden: Ontologies are not yet SMEable, nor Browse-able, nor interactively created or edited visually. Adoption would be accelerated with something intuitive and accessible.

[12:52] Douglas R. Miles: When we see these as Tracks and not themes these are very great

[12:53] ToddSchneider: Bobbin, 'SMEable'??

[12:54] Douglas R. Miles: SME = Subject Matter Expert

[12:54] MikeBennett: Neuro Symbolic - check out Igor Aleksander of Imperial College e.g Aleksander and Mmorton, mid 90s.

[12:54] ToddSchneider: Ram, Could 'Learning Ontologies' be modified to have a system 'learn' to use an ontology (i.e. to modify itself)?

[12:55] ToddSchneider: Bobbin, I assumed SME referred to Subject Matter Expert, but what does the 'able' suffix refer to?

[12:55] DrRaviSharma: neurosymbolic is combination of Neural networks and ?

[12:56] DrRaviSharma: welcome Todd

[12:56] MikeBennett: Melanie Mitchell - spoke today at ER20 - possible speaker?

[12:57] DrRaviSharma: Todd you mean self learning ontologies?

[12:57] Douglas R. Miles: Bobbin, SMEable = Accessible to the SME?

[12:58] DrRaviSharma: Smart materials fix cracks and fractures but how do ontologies? One way would be design patterns matching etc.

[13:00] ToddSchneider: Bobbin, SME's should not be left alone in developing an ontology.

[13:01] Douglas R. Miles: "Tools that Ontologists may use to work together" By Douglas

[13:02] Douglas R. Miles: (For Ken)

[13:02] ToddSchneider: Gary, are your slides available?

[13:02] MikeBennett: Neuro symbolic computing - also requires that we know how to characterize ontologies so we know the right kind to use. So it fits the Ontology Engineering remit

[13:02] ToddSchneider: Ravi, no. I had self-modifying systems.

[13:02] DrRaviSharma: thanks Ram for clarifying neural networks and ontologies

[13:03] BobbinTeegarden: @Doug yes, something you could easily teach a SME (non-technical domain person) to do, like they now develop class diagrams on the fly to show their 'data'/information. Something like a mind map or class diagram that underneath generates good ontologies...

[13:03] Mark Underwood: Roughly speaking, my suggestion was related to Design patterns for Ontologist (not "ontology") collaborations with SME's / domain experts

[13:03] DrRaviSharma: Mike - thanks for many inputs, will hopefully catch them in recording.

[13:04] BobbinTeegarden: @Mark drag and drop ontology design patterns would be a great addition

[13:05] DrRaviSharma: Todd - self modifying against ? a pattern, standard design or foundation?

[13:07] DrRaviSharma: Todd we have perhaps not yet reached a stage of evolution where automatic ontology development happens like mutation etc?

[13:10] BobbinTeegarden: KADS Patterns were little semantic graph patterns that can be combined in many ways to create 'thoughts' -- micro-semantic patterns, reuseable in the small

[13:10] DrRaviSharma: Ken I would be interested in workshop invite.

[13:11] DrRaviSharma: Ravi is copying from Zoom chat

[13:11] BobbinTeegarden: Go look again at KADS out of Euro. Esprit Consortium from 90s...

[13:11] DrRaviSharma: [Post by Gary Berg-Cross on the Zoom chat]

Harmonizing Semantic resources topic

  • Builds on recent summits.
  • Think of the semantic gradient. In the past dictionaries and lexica were powerful building blocks but the success of the Semantic Web since we started has produced a large quantity of available "structured data" that can be leveraged for various purpose such as KG apps.
  • But we need all of them and they are diverse.
  • Ontologies have some advantages but people are still building controlled vocabaularies.
  • ML can now extract some knowledge from text and data sets to help build semantic resources.

Modern Ontological Engineering topic

  • Ontologies are products of Ontological Engineering.
  • The practice has changed in part because the products have changed and both are more mature.
  • Example ontology modules vs domain ontologies or reference ontologies
  • How are KGs built in comparison to ontologies?
  • Are there different standards?
  • Relations and best practices are not well understood.

Resources

Previous Meetings

Next Meetings

... further results