ppy/chat-transcript_edited_20130919b.txt


------
Chat transcript from room: summit_20130919
2013-09-19 GMT-08:00 [PDT]
------


[8:16] PeterYim: Welcome to the 

 = OntologySummit Theme: Community Brainstorm Session - Thu 2013-09-19 =

Session Co-chairs: Professor MichaelGruninger & Mr. PeterYim

Topic: Brainstorm Session on the OntologySummit2014 Theme: Community Input & Planning


Logistics:

* Refer to details on session page at: http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2013_09_19

* (if you haven't already done so) please click on "settings" (top center) and morph from "anonymous" to your RealName (in WikiWord format)

* Mute control: *7 to un-mute ... *6 to mute

* Can't find Skype Dial pad?
** for Windows Skype users: it's under the "Call" dropdown menu as "Show Dial pad"
** for Linux Skype users: please note that the dial-pad is only available on v4.1 (or later or the earlier Skype versions 2.x,)
   if the dialpad button is not shown in the call window you need to press the "d" hotkey to enable it.

Attendees: MichaelGruninger (co-chair), PeterYim (co-chair), AliHashemi, BobSmith, HensonGraves, JackRing, KenBaclawski, LamarHenderson, LeoObrst, MatthewWest, MikeDean, Richard Martin, StefanoBorgo, ToddSchneider


  = Proceedings ==

[8:33] PeterYim: == MichaelGruninger & PeterYim starts session by providing an overview - see intro slides at: http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2013_09_19#nid3XST

[8:42] List of members: AliHashemi, Bob Smith, HensonGraves, KenBaclawski, LamarHenderson, LeoObrst, MichaelGruninger, MikeDean, PeterYim, Richard Martin, ToddSchneider, vnc2

[8:43] PeterYim: == open discussion begins ...

[8:44] LeoObrst: How about, a slight merger of topics: Ontologies and Reasoning over Big Data

[8:44] PeterYim: Michael suggested ToddSchneider starts the discussion off ... telling us about his suggestion of the "Reasoning with Ontologies" theme (that he brought up during the OntologySummit2013 postmortem session)

[8:45] PeterYim: @Todd - how do you see this theme interact with the RulesReasoningLP mini-series that is just starting on Ontolog?

[8:47] ToddSchneider: Peter, yes. They overlap.

[8:--] PeterYim: I suggest we capitalize on the buzz in "Big Data" and work on a theme that shows the world how Ontology can contribute to Big Data

[8:48] AliHashemi: I'd be very interested in Ontologies in Hybrid Reasoning Systems.. It is relevant to both Todd and Peter's points.

[8:49] AliHashemi: Given that in many contexts, people assert that reasoning over declarative constructs is not particularly efficient, it would be interesting to see how various people cordon off declarative reasoning, and pass off the reasoning task to some alternative system

[8:49] AliHashemi: I suspect many Big Data problems would invoke such a methodology.

[8:50] LeoObrst: There is a AAAI Fall Symposium on "Semantics for Big Data": http://www.aaai.org/Symposia/Fall/fss13symposia.php#fs04. Also, STIDS 2013 theme is "Semantics Technologies for Big Data": http://stids.c4i.gmu.edu/.

[8:52] HensonGraves: Metadata and its generalization to ontologies for managing large data is one promising area where reasoning is applicable for finding and drawing conclusions about data

[8:53] ToddSchneider: Another aspect of big data is search and semantic search.

[8:57] MichaelGruninger: I really like the idea of a Summit that brings together the Semantic Web community with the formal and applied ontology community

[8:57] PeterYim: Leo / Michael: ... the merged "merger of topics: Ontologies and Reasoning over Big Data" (plus Ali's suggestion on "hybrid reasoning") provides a theme that will bring the ontology community and the semantic web community together too

[8:58] LeoObrst: The Linked Data community typically addresses the lower range of the Semantic Web, i.e., RDF + URIs/IRIs. That too is "big" distributed data.

[8:58] AliHashemi: A common complaint from the ontology community is that sem web vocabularies don't often have explicit semantics. In many cases, these semantics are interpreted by other tools, because ontology reasoning tools are not perceived as being adequate to the task.

[8:59] AliHashemi: Elaborating on these rationales, and identifying why and how the reasoning is passed to external processes would be informative

[9:00] AliHashemi: Isn't there a danger at casting too broad a scope for the theme? It can detract from generating something actionably useful for the community(ies)

[8:59] ToddSchneider: Peter, what's the schedule for the Rules, Reasoning, LP meetings?

[9:00] PeterYim: @Todd - see: http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2013_09_12#nid3XF2

[9:01] PeterYim: mini-series sessions:    (3XF2)

    R1. RulesReasoningLP: Survey and Introduction [LeoObrst, BenjaminGrosof]    (3XF3)
       session date: Thu 2013.10.24    (3XF6)
    R2. Concepts and Foundations of Logic Programming and Semantic Web Programming [LeoObrst, PascalHitzler]    (3XF7)
        session date: Thu 2013.10.31    (3XFA)
    R3. Rule Standards: Common Logic, RuleML, and RIF [HaroldBoley, AdrianPaschke, MikeDean]    (3XFB)
        session date: Thu 2013.11.21    (3XFE)
    R4. Guide to Reasoning Applications Development and Cases [HensonGraves, KenBaclawski]    (3XFF)
        session date: Thu 2013.12.19    (3XFI)

[9:02] LeoObrst: The "standard" approach to Semantic Web / OWL + logic programming (with rules) reasoning has been explicitly hybrid, i.e., some approaches in logic programming use two parts to any rule, a description logic part and a standard logic programming part. Other approaches do try to merge these, e.g., Grosof's work, etc.

[9:03] AliHashemi: Can we identify some of the pain points @ the interfaces between the two communities?

[9:03] AliHashemi: My impression is that the Semantic Web community stays away from more formal ontologies because (a) they're hard, (b) suspicions about the value of reasoning over the ontologies

[9:03] AliHashemi: *i could be wrong*

[9:04] HensonGraves: upper ontologies are quite useful for describing large data sets

[9:06] ToddSchneider: Henson, but the descriptions (based on a foundational ontology) usually don't support the detailed analysis that users expect.

[9:08] HensonGraves: Todd, the upper ontologies give a good, common place to start, they may have to be enriched

[9:05] ToddSchneider: What about the something along the lines of creating ontologies from big data?

[9:05] AliHashemi: If the idea is to try to bridge the gap, I would suggest we do some "market research" and identify what the real pain points between the two communities are

[9:05] AliHashemi: and perhaps focus in on these issues.

[9:06] ToddSchneider: Ali, which communities?

[9:06] AliHashemi: for @12:05 - communities would be Formal Ontology and Semantic Web

[9:09] ToddSchneider: Should we spin the notion as 'big linked data':)?

[9:09] PeterYim: when we are doing an "internal-facing theme" we want to be *very focused* ... but when we are doing an "external-facing" theme, having a wider scope can actually be useful, and allow us to be more inclusive, get more people engaged, and so on

[9:10] Richard Martin: How does the topic proposal for the two communities fit with the topic of Semantic Federation?

[9:12] anonymous morphed into MatthewWest

[9:15] MatthewWest: Apologies for late arrival.

[9:12] MikeDean: JimHendler named director of new RPI Institute for Data Exploration and Applications (IDEA):  http://prod3.server.rpi.edu/oth/update.do?artcenterkey=3060

[9:12] AliHashemi: Well, why does Jim think that deep semantics are not fruitful?

[9:12] AliHashemi: And why do we give the impression that lightweight semantics are not enough?

[9:19] JackRing: Beyond today's session focus on the [Ontology community + Semantic Web community] is the N other communities who do not have necessary, sufficient and efficient ways of understanding one another. Is it time to address the larger problem?  Until we respond to the issue of semantic federation we will not serve the "marketplace"

[9:26] JackRing: This may be too far off the wall but please consider why the Morse Code was embraced by many people world-wide whereas Esperanto exhibited far less market standing.

[9:33] JackRing: As the bumper sticker says, "Ontologies Happen" or is there a way to precipitate an ontology representative of some field of discourse? The decisions and choices that arrive at an ontology are quite similar to those that arrive at a model of most any system, describing its intended purpose along with what it IS, KNOWS and DOES. A large vote for how to arrive at an ontology.

[9:20] AliHashemi: IMO, the hybrid reasoning approach would help bridge these multiple communities. In each community, what an ontology can offer is limited by a variety of factors, and how the ontology operates in the larger systems is dependent on the perception of the community of the limits of ontology, and the passing off of erstwhile reasoning tasks to other system components.

[9:20] BobSmith: The Financial Industry BUSINESS Ontology (FIBO) effort seems rich with potential facets to discuss

[9:23] PeterYim: @RichardMartin, JackRing, BobSmith, LamarHenderson - are you alluding to the use of "ontological analysis" as the means whereby we can reach out to a lot of other communities?

[9:24] LeoObrst: "Big data" is a bit of a misnomer. Typically what is meant is vast uninterpreted, barely interpreted, or disparately interpreted data. Hence the usual use of quantitative methods (statistical, data-mining, etc.) But there is other big data, such as in e-science, the financial community, where there is a need for finer semantic interpretation.

[9:24] HensonGraves: Peter, yes i agree

[9:30] ToddSchneider: Use of ontology for managing big data?

[9:31] HensonGraves: Todd, as I was saying earlier ontology is very useful for managing big data. this is using an ontology as a generalized version of metadata

[9:32] ToddSchneider: If the theme is ontology development, do we then have a ready set of speakers (i.e., the upcoming book contributors?:)

[9:33] MichaelGruninger: Theme 1: Semantic Web and Formal Applied Ontology

[9:33] MichaelGruninger: Theme 2: Semantic Federation

[9:34] MichaelGruninger: Theme 3: Development of Ontologies

[9:34] RichardMartin: Assuming that 1) a very specific domain ontology is possible for each of several related domains, 2) how do we use ontological tools to communicate amongst those communities?

[9:34] MatthewWest: The challenge with Big Data and Linked Data is that it is just data, and the problems are the same problems we have discussed before, mostly about meaning and integration. However, engaging with the latest buzz words obviously has some merit. I would also be interested in looking at some of the "What" we identified in this year's communique, and developing it into a "How to" methodology for some of those tasks.

[9:35] PeterYim: Theme 3a: Development of Ontologies within and across domains

[9:35] ToddSchneider: If the theme is development of ontologies, then I would expect the outcome and the communique to provide a paradigm for developing ontologies.

[9:35] HensonGraves: development of ontologies as a way of bridging different communities, and communities such as construction which have to broaden their conversation

[9:36] JackRing: If we pursue Theme 3 then Semantic Federation becomes a facet of the design process. Then we notice that Big Data is a pile of needles in which there may be several straws belonging to several systems.

[9:37] LeoObrst: Semantic Web and Formal Applied Ontology for Big Data

[9:38] PeterYim: @LeoObrst and All: how about - Theme 1a:Semantic Web and Formal Applied Ontology for Big Data

[9:38] LamarHenderson: We need to define big data.

[9:38] MatthewWest: Rather we need to discover how Big Data is being used.

[9:43] RichardMartin: How do we craft a domain ontology to maximize the use of reasoning methods across domains?

[9:44] LeoObrst: Potential tracks: 1) definition of big data and kinds of semantic interpretations needed, i.e., use cases. 2) combining ontological reasoning and lightweight SW reasoning over big data. 3) Combining quantitative methods with ontological reasoning over big data.

[9:45] JackRing: 2 and 3 are not the same. 2 presumes entities already exist and must be rationalized whereas 3 creates the entities

[9:47] AliHashemi: For what it's worth, while I believe reasoning could be a single track in the Semantic Web + Formal Ontology track is viable

[9:47] ToddSchneider: How about the application of ontology analysis in big data and the semantic web as a theme?

[9:47] AliHashemi: It is a distinct theme, because reasoning in Systems vs reasoning in Semantic Web vs reasoning in Big Data have separate components

[9:48] AliHashemi: and could be individual tracks each

[9:48] LeoObrst: Also, a topic like "Semantic Web and Formal Applied Ontology for Big Data" kind of teams up the two communities to collaboratively address "big data" issues.

[9:48]MatthewWest: It would be better to pick something specific rather than try to combine everything.

[9:48] HensonGraves: got to go, by

[9:49] AliHashemi: I've got to run. Bye.

[9:50] LeoObrst: I must go too.

[9:52] PeterYim: summarize (title and one paragraph): Theme-1 - MichaelGruninger; Theme-2 - RichardMartin; Theme-3 - ToddSchneider -> Michael

[9:53] PeterYim: Michael will then post to [ontology-summit] list and we discuss these theme there further (with the rest of the community)

[9:53] PeterYim: great session ... thanks everyone!

[9:54] PeterYim: -- session ended - 9:52am PDT --

------