Actions

Ontolog Forum

Session Synthesis
Duration 1.5 hour
Date/Time Apr 04 2018 16:00 GMT
9:00am PDT/12:00pm EDT
5:00pm BST/6:00pm CST
Convener KenBaclawski

Ontology Summit 2018 Communiqué Session 1

Agenda

The Communiqué will be organized as follows:

Video Recording

The Symposium will be held on 30 April and 1 May 2018 at NITRD in Washington, D.C.

Conference Call Information

  • Date: Wednesday, 04-April-2018
  • Start Time: 9:00am PDT / 12:00pm EDT / 6:00pm CST / 5:00pm BST / 1600 UTC
  • Expected Call Duration: ~1.5 hours
  • Video Conference URL: https://bluejeans.com/703588230
    • If you have not used BlueJeans before, then connect to the URL above before the meeting time so that the required plug-in can be installed.
  • Chatroom: http://webconf.soaphub.org/conf/room/ontology_summit_2018
    • Instructions: once you got access to the page, click on the "settings" button, and identify yourself (by modifying the Name field from "anonymous" to your real name, like "JaneDoe").
    • You can indicate that you want to ask a question verbally by clicking on the "hand" button, and wait for the moderator to call on you; or, type and send your question into the chat window at the bottom of the screen.
  • This session, like all other Ontolog events, is open to the public. Information relating to this session is shared on this wiki page.
  • Please note that this session may be recorded, and if so, the audio archive is expected to be made available as open content, along with the proceedings of the call to our community membership and the public at-large under our prevailing open IPR policy.

Attendees

Proceedings

[12:07] ToddSchneider: The overall organization of the Communique is planned to be the following:

  1. Introduction
  2. Domain Specific Needs http://bit.ly/2DLK7FY
  3. Upper Ontologies http://bit.ly/2DLaZ9l
  4. Open Knowledge Network http://bit.ly/2DKnXUS
  5. Integration and Interoperability http://bit.ly/2DOiBI9
  6. Systems Engineering http://bit.ly/2DNoEwz
  7. Representation of Context http://bit.ly/2EC9Z6S
  8. Summary of Issues http://bit.ly/2IJFFKe

[12:08] KenBaclawski: The link to the section of the Synthesis meeting page with the summaries by Ravi Sharma: http://ontologforum.org/index.php?title=ConferenceCall_2018_04_04&oldid=24803#hid2F1

[12:13] ToddSchneider: Is the a theme or multiplicity of themes (other than context) that tie the planned sections together?

[12:17] ToddSchneider: How shall, or should, 'agency' be addressed?

[12:18] Gary Berg-Cross: I would suggest adding a Findings section to our Outline before the Issues section. And we need a references section.

[12:18] ToddSchneider: Can the communique suggest a start at ways to qualify 'context' for operational purposes?

[12:23] RaviSharma: Ken said we should emphasize on issues. But is it also not that first we try to summarize what the domain-context depiction is?

[12:24] RaviSharma: Notes: Domain-context summary in one short para is desirable.

[12:25] RaviSharma: before we enumerate issues, yes blog can contain summary a good idea but then you have to allow links in the Communiqué.

[12:26] BobbinTeegarden: @Ravi could you put your slides on the website?

[12:27] RaviSharma: @Bobbin, they are at link http://ontologforum.org/index.php?title=ConferenceCall_2018_04_04&oldid=24803#hid2F1

[12:27] RaviSharma: the PDF file.

[12:28] Gary Berg-Cross: Our blog has a preliminary outline of OKN material which may also be useful for the Summit Communique. Material is roughly organized as follows:

  • Relation of OKN in the context of discussions from prior Ontology Summits
  • An overview of Extraction and Knowledge Graph Building
  • Contextual Examples of interest to OKN
  • Preliminary findings
  • Challenges and issues to investigate

[12:29] RaviSharma: Notes: Mike Bennett will also summarize the upper ontology and put on blog.

[12:29] Gary Berg-Cross: In the OKN blog we follow the guidance laid down for the Summit by John Sowa in his writings about context. As he notes there are four kinds of context involved in language understanding (which is relevant more widely in other cognitive products):

  • the text or discourse;
  • the situation; reflecting in part that is a context is a context of something. It may be text as above but also physical and mental realms.
  • common background knowledge; and
  • the intentions of the participants.

Some connection to these different types of context are noted in the context of an OKN discussion. So open knowledge may be created/extracted from discourse on the web, from some particular structural/format situation, and interpreted using some background knowledge captured in an ontology or some processing tool and reflecting some intention of an intelligent agent to document information.

[12:29] John Sowa: Context is not part of the ontology. It belongs to the methods for using an ontology/

[12:30] John Sowa: Context is totally dynamic. It changes constantly.

[12:31] RaviSharma: @John, could there be also an ontology that deals with parameters (things) and how they relate to context?

[12:31] ToddSchneider: John, doesn't an ontology embed some 'parts' of a context?

[12:32] John Sowa: Anything in an ontology, a database, or the WWW may be part of the context of any sentence.

[12:32] RaviSharma: @John yes context can be mostly dynamic and can ontologies be dynamic also, like a movie?

[12:32] John Sowa: Todd, everything in the ontology can be part of the ontology.

[12:33] John Sowa: The ontology is static -- at least for any particular problem.

[12:33] John Sowa: But the context is dynamic.

[12:34] John Sowa: An ontology of context would be identical to an ontology of everything.

[12:36] ToddSchneider: John, aren't there 'parts' of a context that change more slowly than others (e.g., stability of protons)?

[12:36] RaviSharma: Notes: Mike said about: granularity, listed items for context, context outside has to influence ontology, pragmatics, everything is context, context and perspective, ref ontology like FIBO is broad, others can be narrow, good enough for engineering, symbols, intentions,

[12:36] RaviSharma: 1. scoping

[12:37] Jack Ring: We should mention what is not an ontology

[12:37] RaviSharma: understanding, why perspective, items of interest, outside to inside, gallery of ontology, perspectives begin and end.

[12:37] Cory Casanave: John: re: Context is not part of the ontology... Given some text, other text is applicable for interpretation - supporting text, this seems to be the context. Some supporting text may be "included", directly stated in the same ontology. Others "indirect context", assumed for interpretation.

[12:39] Jack Ring: What makes anything an ontology? What is it just before it becomes an ontology

[12:40] pfps: One could think of a higher-level ontology that has contextualized parts.

[12:41] Gary Berg-Cross: @pfps Yes, that seems a useful approach to this.

[12:42] Cory Casanave: John: "People do not discuss context": When people say what they are assuming or their perspective they are communicating their context.

[12:42] pfps: One can even characterize Cyc's microtheories as both a reasoning control mechanism and as part of Cyc's overall ontology.

[12:43] BobbinTeegarden: If you back up (abstract out) of the ontology to the context of an ontology, it becomes a bigger, more abstract ontology -- but still 'an ontology'

[12:44] BobbinTeegarden: An Ontology which contains the subject matter ontology plus the context of the subject matter ontology...

[12:44] MikeBennett: Exactly

[12:46] pfps: There is the question of whether the consequences of an ontology that are related to the ontology itself are part of the ontology. Some Cyc microtheories have as their main consequence the conflation of two different kinds of things, for example conflating the organizational part of a nation with its geographical extent.

[12:46] RaviSharma: Notes: John said context brings info from anywhere but context is about method by which you use or operate on ontology. He also said ontology of context is a minor part, context belongs to method of reasoning, Mike said customer relationships in Bank, John said ontology of database is mostly fixed but not a part of reasoning, Bank does not have context ontology, but analysis of a Question for bank can use ontologies in use in Bank, David Whitten wants to know Cyc Lenat work, context and microtheories, John said yes meta level, lot of microtheories, reasoning at meta level, David said that variables are not part of ontology (?) John said what is meta Level, reasoning over database elements, microtheories subdivide the space in parts, microtheories help in organizing ontologies.

[12:46] Cory Casanave: Perhaps we have just identified one of the open issues!!

[12:47] RaviSharma: Notes: John emphasized that context is outside ontology and is related to reasoning.

[12:48] RaviSharma: Notes: John said this has to be part of communique'.

[12:48] John Sowa: Ravi, no ontology anyone has ever developed is dynamic. it may change over time, but very slowly.

[12:50] John Sowa: FUNDAMENTAL PRINCIPLE: Context is part of the communication process and the reasoning process.

[12:50] ToddSchneider: John, would it be sufficient to use the Tarskian view to show how context is external to the use/application of ontologies?

[12:51] RaviSharma: @John Thanks, but how we depict ontology of earth images, moving screens such as satellite imagery over orbits? of course context is changing with terrain, orbit parameters and sunlight angle, spectral filters, etc?

[12:52] pfps: If you think of the ontology as your outermost organizational principles then it shouldn't change quickly. If, however, you think of the ontology as your current organizational principles, then that ontology could change quite rapidly in response to context switches.

[12:52] John Sowa: Todd, Tarski said nothing about context. He was talking about the entire ontology as a single theory.

[12:53] pfps: I would think that there should be a current pointer to the state of OKN so that readers of the communique can find out what is going on with OKN.

[12:53] John Sowa: When you organize an ontology in microtheories, it helps to find relevant parts during the reasoning process.

[12:53] RaviSharma: Notes: Gary described the blog page and summarized.

[12:54] John Sowa: pfps, that pointer would be a dynamically changing selector for finding which parts of the OKN is relevant at the moment.

[12:55] pfps: Is there anything in OKN that is not fuzzy??

[12:55] ToddSchneider: John, I was alluding to context as constraint on interpretation.

[12:56] pfps: @john: I would expect it to be a pointer to some document showing the current state of OKN, which of course could contain pointers to whatever is happening there.

[12:56] John Sowa: During any reasoning or communication process, the ontology is static.

[12:56] Ram D. Sriram: Regarding John Sowa's point, I believe that we need to point it out in the communique. It is interesting that we are at the end of the summit and still arguing about context. I think the other view of embedding contexts into ontologies should be argued more convincingly. I think both these views can be addressed properly.

[12:56] RaviSharma: Notes: Ram said to Gary that it was a great job and beginning, overall thought that OKN is gaining attention, Guha and others work, need OKN, how exactly to be represented, in relation to Communique' we need to summarize. Agreed with John's point, but context in ontology is also to be mentioned.

[12:57] John Sowa: The context is that part of the static ontology or the database that is being used in the current step.

[12:57] Gary Berg-Cross: Ravi: Do you have a summary of interoperability?

[12:58] Ram D. Sriram: @Gary: Totally agree with you for the need of a good example

[12:58] RaviSharma: Notes: Jack Ring said it is imp for users and builders on ontology to be discussed in Systems Engineering, problems exist because we do not have many good ontologies for Developers.

[12:58] John Sowa: Ram, How on earth can you embed a dynamically changing context into a static ontology?

[13:00] MikeBennett: That is one (extremely important) meaning of the word 'context/. Another is when people talk about the interpretation of some specific symbol in terms of roles, systems, temporality, location etc. - things that are often implicit in a given stand alone data model (or even a stand along OWL application) but can be explicitly represented within a broader reference ontology (or any ontology). If we don't call those context, what would we call them? What you call context can as easily be labeled as 'perspective'. I don't think we should be taking strong positions on the use of words, but rather focus on the shape of the problem.

[13:00] RaviSharma: Notes: Gary said Ken will attempt to give it a try, Jack Ring said antidote for Polio, Salks' research theories will be related? Gary asked for email on this thought.

[13:01] John Sowa: Mike: Imagine the entire ontology and all its microtheories written on white walls of an enormous room.

[13:01] Ram D. Sriram: @John: I am not saying I know it. Currently, my notion of context is similar to yours. All I am indicating is that there seems to be an argument for it and we need a good use case to show how "to embed a dynamically changing context into a static ontology."

[13:02] John Sowa: Ram, I'm glad that you agree. I believe that Mike B. would also agree, when I finish the analogy.

[13:03] Ram D. Sriram: @john: "When is said "argument for it" I meant "the other view of dynamically changing context"

[13:03] janet singer: I agree with John that we need to clarify what is meant by context and ontology up front for the communique to be coherent. I agree with Mike that this requires a few key context-ontology relationships to be identified for the general case.

[13:03] MikeBennett: @John that's just the sort of thing I would do. Then the room is the Sowa-Context and the classes that represents abstractions such as mediating things, time, system, process etc. etc. are kinds of Bennett-context. One can never do away with the room - the perspective or context in which the ontology is referred to.

[13:04] John Sowa: Given the big room with white walls covered with the entire ontology. Then think of the context as a bright light that is shining on any part of the ontology that is being used at the moment for that particular step.

[13:04] RaviSharma: @all: my view is that John has described most useful and likely to be most relevant description of context relating to method, and as others said, the context in ontologies is a less prevalent use of context? Comments are welcome in relation to communique'.

[13:05] John Sowa: Whatever the light is shining on at any moment is in the context.

[13:06] John Sowa: The same is true of the database, the OKN, or anything else. Whatever the light is shining on is in the context at that moment.

[13:07] MikeBennett: Here is a link to the slides that I forgot to display: https://s3.amazonaws.com/ontologforum/OntologySummit2018/UpperOntology/Track+B2+Outcomes--Context+and+Upper+Ontology--Bennett.pptx

[13:07] MikeBennett: Ken where on the session page do you want these?

[13:07] RaviSharma: Notes: Cory mentioned and summarized integration and interoperation and referred to Hans Polzer and others points in discussion, ontologies with textual relationships in them, capture some of those in higher order relations. Interpretation context is what is used,in analyzing relationships.

[13:08] MikeBennett: @John great analogy - let's use that in the Communique as it shed light (pun intended) on what we are talking about.

[13:11] Gary Berg-Cross: I can think of the "light" idea to represent human attention and its focus.

[13:11] Cory Casanave: Time is a dynamically changing context, but we can model it. Being on the surface of the earth changes slowly but impacts assumptions.

[13:12] John Sowa: Mike and Gary, I'm glad that we are resolving the issues.

[13:13] BobbinTeegarden: @John "Given the big room with white walls covered with the entire ontology. " but that big room with white ... etc is a context unto itself.

[13:14] John Sowa: Mike, For a single definition of context, start with the definition: the immediate context of a sentence in a document or a discourse is the surrounding text (or dialog).

[13:14] Ram D. Sriram: Here is an example from Jain philosophy that might be relevant to our discussion. The Jain philosophy believes that "Absolute knowledge (even if we possessed it) could not properly be communicated using limited concepts, all claims should be seen as relative and provisional, and apparent contradictions may simple be the result of differences in perspectives. Suppose I look at a tree, I may say: 1) There is some useful shade or shelter; 2) That would be good to climb; 3) There is some useful timber; 4) It is an Oak; 5) It has grown since I saw it last; 6) It will have to be felled. Each of these perspectives is valid, whether it is concerned with past, present or future, and whether the tree is seen in itself (and categorized in terms of species) or simply viewed as a commodity or source of entertainment or comfort. Each has an element of truth, but each is limited by its particular perspective. Of course, I may be mistaken -- the tree is a beech rather than an oak -- but that may not invalidate the other statements about it." The above quote is from Mel Thompson's book on "Understand Eastern Philosophy."

[13:14] Gary Berg-Cross: The big room]s walls are containers.

[13:14] janet singer: Moving light is a good metaphor but we need to generalize it to interpretation events beyond the human

[13:14] RaviSharma: Notes: Systems Engineering- Jack summarized how people design develop ontologies, but not dealt with issue why 50% of Sys Engineering projects fail, need better Air traffic, manufacturing, etc, devise tech systems engineering solutions, process for doing vs efficacy measures are challenges. People do not make sure they understand the problem, they do not address the issues or solution approach. they often dealt with what they knew. This group Stafford Bear(?) said about intention of user.

[13:15] John Sowa: The extended context is any background knowledge that is necessary or helpful to supplement the immediate context.

[13:16] Gary Berg-Cross: X is part-of-system Y is a particular "contextual" part relation.

[13:16] MikeBennett: @John 13:14 that's precisely how I have always seen it, which is why I would want as many of those things as possible within the ontology. Some will always be outside and be the context in which the whole ontology is seen (the one you are talking about), those one has managed to include are the context, within the ontology, of any given symbol in that ontology. This was the gist of the slide I belatedly remembered to display.

[13:16] RaviSharma: Notes: Jack - requires ontology of problem situation, what might be done? Sys Engineering does not apply test for purpose.

[13:16] John Sowa: That background knowledge can come from anywhere: the ontology, the database, the www, or the physical situation (time, place, other people, things, weather...).

[13:17] MikeBennett: @Ram 13:14 I think I would like to steal that for our blog page.

[13:17] RaviSharma: Notes: Jack Requirements vs purpose. Janet said arms length objectives: deception, perception, communication, ontology specific enough, esp for era of autonomous systems.

[13:17] BobbinTeegarden: So is there an explicit context, and an implicit context apart?

[13:17] John Sowa: Mike, I don't understand what you mean by "wanting those things in the ontology".

[13:18] John Sowa: Do you mean that you are shining the light on them?

[13:18] MikeBennett: @Bobbin yes that's how I see it. Many things that are implicit for a given data model e.g. a customer relationship database, can be made explicit as classes in a broader reference ontology, that can (therefore) be used in integration, reporting etc.

[13:19] John Sowa: Ravi, all those things that Jack and Janet are saying are part of the reasoning process.

[13:19] MikeBennett: @John I was thinking in terms of having classes of 'Thing' for time, place etc., which would not be needed for a stand-alone application (implicit) but can be included in the reference ontology (with properties pointing to it) in the broader ontology.

[13:20] John Sowa: But I agree that it would be useful to have an ontology about reasoning methods -- including ways of using the ontology.

[13:20] BobbinTeegarden: @John As soon as you shine the light on them, aren't you making that part of the context a part of the ontology?

[13:20] MikeBennett: @John the shining a light one is an interesting idea - can we define dynamically created application ontologies from a broader reference ontology, by considering the light?

[13:20] John Sowa: But an ontology about reasoning methods is a *metalevel* ontology.

[13:21] Cory Casanave: What is wrong with metalevels?

[13:21] BobbinTeegarden: @John How many levels of metalevels are there?

[13:21] John Sowa: In the previous question about Tarski, were you taking about Tarski's hierarchy of metalevels?

[13:21] RaviSharma: @all: Three examples: Apollo Systems Engineering, ERP of manufacturing Cars and its automation by forward solution architecture and also Long range procurement programs in DoD are some examples of systems engineering and solution that worked 30-50 years ago do not work because of advances in Technology, purpose and changing paradigms.

[13:21] Mark Underwood: @Gary A context machine obeys certain constraints, some of which model human attention

[13:22] John Sowa: If so, then Tarski's metalevels would be relevant: The ground level (what's on the walls of the room) would be static.

[13:22] Gary Berg-Cross: John noted an immediate context and an extended one. This distinction helps with part of things being conflated in this discussion.

[13:23] Cory Casanave: Have to run...

[13:24] John Sowa: The metalevel would address the room and the searchlight that shines on parts of the (static) ground level.

[13:25] BobbinTeegarden: @John ... but if your perspective is the entire setup including the light shining on parts etc, that is a (wider) context also

[13:25] John Sowa: Any ontology about context would be a metalevel ontology that talks about how the object level info may be used.

[13:25] John Sowa: Bobbin, instead of the word 'wider', use the word 'meta'.

[13:26] Ram D. Sriram: @MikeBennett: "species in the quote is actually species" I am sure you can sue the quote provided it is attributed properly to Mel Thompson's book.

[13:26] BobbinTeegarden: @John, as long as 'meta' has levels, is holonic

[13:26] John Sowa: Tarski talked about a ground level, a metalevel, a metametalevel...

[13:27] Gary Berg-Cross: @John, I take it that this meta-level could include info such as some provenance information about the things in the object level.

[13:27] BobbinTeegarden: @John infinite in both directions: meta levels, and micro (decomp) levels?

[13:27] John Sowa: Every metalevel includes access to all the levels beneath it all the way down to the fixed ground.

[13:28] John Sowa: Tarski assumed a fixed ground.

[13:28] Ram D. Sriram: @Ravi: Systems Engineering (SE) is a pretty big topic. We will probably need an entire summit to discuss the role of ontologies in SE.

[13:29] BobbinTeegarden: @John but implicit order implies no fixed 'ground', but decomposable to smaller contexts... no?

[13:29] John Sowa: Bobbin, the decomposition of any ontology into microtheories is a complicated, slowly changing process.

[13:30] Mark Underwood: @Ram - Book topic as well. Anyone else interested in collaborating on a book or book-like endeavor?

[13:30] Mark Underwood: @Ram - last effort was around 5 year back

[13:30] RaviSharma: Notes: Janet said metalevel has to be addressed as discussed by John, and others. Ken asked Janet and Jack try to capture these in blog, Jack wrote a paper, we can revisit. Jack needs to revisit, learn in John's terms then a short and informative comment can be put in Communique'

[13:30] BobbinTeegarden: @John is a way to say that 'contexts are holonic'?

[13:31] Ram D. Sriram: @Mark: I believe so. Just underscoring

[13:33] RaviSharma: Notes: Jack said integration of ontologies required because of incomplete systems engineering. Meeting set up Friday 13th at 1 PM EST with John Sowa.

[13:35] MikeBennett: I've made a start of the Upper Ontologies blog by putting the slides to it and a summary. Need to do more.

[13:36] janet singer: Having network problems in BlueJeans, but Friday 1 EST works

Resources

Video Recording

Ontology Summit Synthesis Page

Ontology Summit Symposium

Previous Meetings

... further results

Next Meetings