Actions

Ontolog Forum

Revision as of 04:17, 19 November 2019 by imported>KennethBaclawski (→‎Ontology Summit 2020 Topic Discussion)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

Session Topic Discussion
Duration 1 hour
Date/Time June 26 2019 16:00 GMT
9:00am PDT/12:00pm EDT
5:00pm BST/6:00pm CEST
Convener Ken Baclawski

Ontology Summit 2020 Topic Discussion

Agenda

Open-ended discussion of the topics for summer meetings and the next summit.

  • Revisit some of the previous summit topics
  • Ontology for the summits

Ontology Summit Topic Suggestions

  • For major domains Create hype-cycles based on Ontology Use Cases and Adoption of ontology applications and techniques – document benefits of using ontologies.
  • Audio-Visual (graphic) and NLP representations of knowledge enhanced by ontologies
  • Meaning, understanding, knowing - using ontologies
  • Interdomain and subdomain vocabularies/terms and ontologies as facilitators in cross linking them
  • Cloud architectures with ontologies
  • Graphics and ontologies example ref: Graph-based Ontology Summarization: A Survey [1] also [2]
  • An Overview on Visualization of Ontology Alignment and Ontology Entity (China Euro Conference ECC 2018)

Conference Call Information

  • Date: Wednesday, 26-June-2019
  • 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://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

[12:10] Ken Baclawski: The Video Conference URL is https://zoom.us/j/689971575 which is the same as the one we have been using. I managed to extend the expiration date of this URL so that we can continue to use it. The chat room has changed but the zoom meeting room has not.

[12:17] RaviSharma1: Ravi has 2 posts for Ken to consider

[12:17] MikeBennett: You do see that 'Old' and 'New'; 'Right' and 'Wrong' have no practical semantics. Unless one somehow memorizes these things. Pragmatically you are either in the same meeting that Ken has started, or you are in a different one.

[12:17] RaviSharma1: Ravi Sharma June 25, 2019 on Communique Ontology Summit 2019

Statistics and Probability are intertwined and embedded.

The usual temporal and superficial distinction between probability and statistics is not accurate, because given the ensemble (data or observations - that are generally statistical, often random, and lead to a set of multivariate distribution) when we take a data-point and classify it, the probability is based on decision rules such as Bayes, and it indicates the bucket to which a given data-point in the ensemble belongs.

Hence uncertainty is inherent in nature, whether we are doing prediction (erroneously sometimes call future, it can be for the present dataset also), and thus Next AI wave will be MPE (Most Probable Explanation) or will use learning based on statistics and often depend on uncertainties in nature.

Financial - Other areas of financial transactions such as UNCEFACT, XBRL, processes replacing SWIFT, BASAL, are where explanations will be part of the content of Automation (using AI) and ontologies and to enable eCommerce.

Medical - Multiple domains exist in Medical areas e.g. Genetics, providers, specialists, triages, nurse and technical treatment practices, clinical, pharmaceutical, surgery, physiology etc. There is differential in domain vocabulary and terms and in translation and meaning equivalence, and in many of these the Explanations embedded ontologies and AI solutions can reduce ambiguities and make inter-domain interoperation more seamless.

[12:18] RaviSharma1: Ravi Sharma June 25, 2019 Ontology Summit 2020 - suggested additions to potential topics

For major domains Create hype-cycles based on Ontology Use Cases and Adoption of ontology applications and techniques document benefits of using ontologies.

Audio-Visual (graphic) and NLP representations of knowledge enhanced by ontologies

Meaning, understanding, knowing - using ontologies

Interdomain and subdomain vocabularies/terms and ontologies as facilitators in cross linking them

Cloud architectures with ontologies

Graphics and ontologies example ref: Graph-based Ontology Summarization: A Survey https://arxiv.org/abs/1805.06051 also https://ieeexplore.ieee.org/abstract/document/8527452

An Overview on Visualization of Ontology Alignment and Ontology Entity (China Euro Conference ECC 201 for example topics Etc.

[12:22] MikeBennett: Semantically the Conclusion is more a Summary or Overview than a Conclusion.

[12:23] RaviSharma1: Ken will put explainable systems, AI with and without ontology

[12:31] RaviSharma1: our sharpness will lie in demonstrating the value of Explainable AI and incrementally better if ontology is used in Explainable AI Systems.

[12:36] MikeBennett: Fair enough - concluding paragraph, not a paragraph setting out what conclusions we drew from the exercise (what we concluded). Heteronyms for 'conclusion'!

[12:40] Ram D. Sriram: NIST RFI on AI and Standards: https://www.nist.gov/news-events/news/2019/05/nist-requests-information-artificial-intelligence-technical-standards-and

[12:41] MikeBennett: Ontology and Standards - Yes. Much needed.

[12:43] MikeBennett: Most standards take a very naive view of how you can allocate meanings to terms. They know they needed it but don't know the pitfalls.

[12:44] Ram D. Sriram: All public comments for above RFI are at: https://www.nist.gov/topics/artificial-intelligence/ai-standards

[12:44] Ram D. Sriram: There are 97 submissions. I will send my summary of these submissions to Ken (and any others interested)

[12:51] RaviSharma1: Ram please send it to me also

[12:52] Ram D. Sriram: Here is what I said. AI systems need to be trustworthy, in particular in those fields where human life is at stake. Trustworthy AI systems should have explainability capabilities. For robust explainability systems one needs ontologies. Standardized ontologies, which are verified and validated, will lead to important innovations (see The Role of Standards in Innovation, R. Allen and R. D. Sriram. See https://www.sciencedirect.com/science/article/pii/S0040162599001043)

[12:58] RaviSharma: janet says Gruber is important.

[13:08] RaviSharma: Revisiting ontology standards

[13:08] RaviSharma: ontology evolution

[13:10] RaviSharma: How about maturity of Ontologies in different domains? related to standards or trends towards standards

[13:12] RaviSharma: Mike's comment about deployable, financial, end use emphasis of ontologies

[13:13] RaviSharma: Marco also mentioned how to tie with Vocabularies terms etc.

[13:15] RaviSharma: Ken suggested we continue this topics discussion next wk

Previous Meetings

Next Meetings

... further results