ppy/summit-symposium_chat-transcript_unedited_20140428a.txt ------- Chat transcript from room: summit_20140428 2014-04-28 GMT-08:00 [PDT] ------ [05:31] PeterYim: Welcome to the = OntologySummit2014 Symposium (Day-1) 28-Apr-2014 = This is the finale of the 4 months of OntologySummit2014 activities - a 2-day workshop and symposium in the Metrpolitan Washington, DC Area hosted by NCO_NITRD. The event is being held at the NSF Board Room. Session details: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2014_Symposium Dial-in details: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2014/WorkshopRegistration#nid484Y Agenda: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2014_Symposium#nid482Q . == proceedings: == . [04:37] anonymous morphed into leo [05:01] anonymous morphed into NSF-venue (chat managed by SteveRay) [05:31] anonymous morphed into PeterYim [05:40] SundayOjo: SundayOjo on line [05:49] anonymous morphed into pfps [05:53] anonymous morphed into coreyleong [05:54] ChristiKapp2 morphed into ChristiKapp [05:55] NSF-venue: We're still getting set up. Just a few more minutes. [05:56] anonymous morphed into Christof Hasse [06:02] coreyleong morphed into CoreyLeong [06:03] anonymous morphed into DanCorbett [06:04] TimFinin: George Strawn of NITRD is giving some welcoming remarks... [06:11] NSF-venue: In general, remember: *7 to unmute, *6 to mute [06:11] SundayOjo: Tshwane University of Technology, Pretoria, South Africa [06:11] ChristiKapp: ChristiKapp, JustIntegration, Inc. Kissimmee, Florida USA [06:11] anonymous1 morphed into Todd Pehle [06:12] anonymous2 morphed into Jesse Wang [06:12] AnneThessen: AnneThessen, The Data Detektiv and Marine Biological Laboratory [06:13] MarcelaVegetti: Marcela Vegetti - National Council for Scientific and Technical Research of Argentina (CONICET) [06:16] anonymous2 morphed into FrankOlken [06:17] CoreyLeong: Corey Leong, Institute of Real Estate Science, Orlando, FL US [06:19] TimFinin: Michael Gruninger is describing the topics of this year's summit [06:19] mosesg: *7 [06:21] NSF-venue: You would mute by pressing *7 on your phone keypad. [06:22] NSF-venue: Sorry, I mean you would unmute by pressing *7 on your phone keypad. [06:23] Gary Berg-Cross: @Amanda what is the twitter handle this year? [06:24] Gary Berg-Cross: Just found it #ontosum [06:25] NSF-venue: @Kuldeep, did you want to say something? [06:26] anonymous morphed into JulioRoa [06:26] PeterYim1: @Kuldeep, can you type out your question on the chat, please? [06:26] anonymous2 morphed into LAPritchard [06:28] anonymous2 morphed into DanCerys [06:31] TimFinin: George Strawn is introducing NSF's Dr. Farnam Jahanian [06:35] NSF-venue: Slide 4 [06:36] NSF-venue: Slide 5 [06:37] NSF-venue: Slide 6 [06:38] NSF-venue: Slide 8 [06:39] FrankOlken: National Science Foundation Critical Techniques and Technologies for Advancing Big Data Science & Engineering (BIGDATA) Program Solicitation NSF 14-543 http://www.nsf.gov/pubs/2014/nsf14543/nsf14543.htm Deadline is June 9, 2014 [06:40] TimFinin: 500M tweets per minute ! [06:40] TimFinin: 500K, perhaps... [ed. Jahanian slide#9 actually says "100,000 New tweets every minute"] [06:41] TimFinin: slide 10 [06:41] FrankOlken: Lead program director for NSF Big Data solicitation is Dr. Sylvia Spengler [06:44] TimFinin1: slide 12 [06:46] TimFinin1: slide 13 [06:48] TimFinin1: slide 14 [06:48] TimFinin1: 15 [06:52] NSF-venue: 17 [06:52] NSF-venue: 18 [06:52] NSF-venue: 19 [06:53] NSF-venue: 20 [06:54] NSF-venue: 21 [06:56] NSF-venue: 22 [06:58] anonymous morphed into ElizabethFlorescu [06:58] NSF-venue: 23 [06:58] Gary Berg-Cross: Insights at the macro level of data depend to some extent on finding patterns in that data. [07:01] TimFinin1: In terms of Daniel Kahneman's Thinking fast and slow paradigm: is big data essentially the former? [07:01] NSF-venue: 27 [07:02] NSF-venue: 28 [07:03] DanCorbett: @Gary - Agreed, but I think that one major point here is that we give up precision and rely more on pattern matching to find new insights. [07:05] anonymous morphed into SteveRay [07:05] NSF-venue: 29 [07:06] FrankOlken: National Science Foundation solicitation: [07:07] Gary Berg-Cross: @Tim [10:01] Well patterns can be discovered by fast processes as opposed to deliberate ones, but some pattern discovery may use more deliberate processes and, for example, require empirical pattern confirmation over time. [07:07] NSF-venue: 32 [07:07] FrankOlken: NSF Data Infrastructure Building Blocks (DIBBs) Program Solicitation NSF 14-530 [07:07] FrankOlken: http://www.nsf.gov/pubs/2014/nsf14530/nsf14530.htm [07:08] SteveRay: Is he talking about tagging the reports, or something more sophisticated, I wonder... [07:08] NSF-venue: 34 [07:09] FrankOlken: Lead Program Director for DIBBS: Amy Walton, Program Director, CISE/ACI, and DIBBS Solicitation Manager, telephone: (703) 292-8970, email: DIBBsQueries [at] nsf.gov [07:09] Gary Berg-Cross: Long-term preservation of data including that for scientists and citizens is a challenge since programs and projects have a limited lifetime. Who funds the preservation investment is an issue. [07:10] NSF-venue: 36 [07:11] Gary Berg-Cross: BTW, sustaining Software investments have a similar issue. [07:13] anonymous morphed into Mark Underwood [07:14] NSF-venue: 41 [07:15] NSF-venue: 42 [07:15] NSF-venue: 44 [07:19] NSF-venue: 50 [07:21] anonymous morphed into David Blevins [07:21] NSF-venue: 55 [07:22] NSF-venue: 56 [07:26] NSF-venue: 65 [07:28] NSF-venue: 67 [07:28] NSF-venue: 69 [07:30] FrankOlken: National Science Foundation Research Traineeship (NRT) Program PROGRAM SOLICITATION NSF 14-548 http://www.nsf.gov/pubs/2014/nsf14548/nsf14548.htm [07:32] mosesg morphed into MosesLesibaGadebe [07:38] Jesse Wang: What are the roles of ontologies in real-world (commercial) solutions? [07:38] Jesse Wang: (Question) [07:41] ChristiKapp: Are you aware of any research on strategies for improving the "data curation process" ? [07:42] Mark Underwood: That Q as about Big Metadata [07:42] Gary Berg-Cross: I'm on the RDA US Advisory Committee if you have questions on RDA. [07:42] TimFinin: research data alliance https://rd-alliance.org/ [07:42] Mark Underwood: NIST Big Data WG is also working with RDA [07:43] Gary Berg-Cross: RDA Plenary 3 in Dublin had a BoF on Semantic Interoperability and we ay develop ijnto an interest group. [07:50] anonymous morphed into BobbinTeegarden [07:58] anonymous morphed into BobbinTeegarden [08:04] TimFinin: @mathew I'll try getting people back in the room, so we can start in a few minutes [08:09] PeterYim1: == MatthewWest presenting the Track-C Report ... [08:10] anonymous morphed into MikeBennett [08:17] Krzysztof Janowicz: is the connection gone or is it me? [08:17] ToddSchneider: Matthew, on slide 10, the use of the term 'model' seems to be used to refer to two different entities. Could you clarify? [08:20] SteveRay: @Matthew: And while you are at it, Im confused about the distinction between knowledge representation language and knowledge modeling language (on your slide 8) [08:21] SteveRay: That last emoticon should have been an "8" and a ) [08:21] AndreaWesterinen: @Matthew, slide 10, about development of "the model and its implementation", I have taken an agile approach to the development of ontologies ... such that you develop an overall architecture, and then the necessary piece parts, in conjunction with its use (which is really important), while still allowing fail-fast. [08:24] Gary Berg-Cross: This issue of what expressiveness to target in an 'ontology' is one reason that general domain ontologies are hard to realize. We may perhaps approach them over a long period of time by working both bottom up from focused modules and maintaining a top down effort with domain people. But funding such efforts is hard. Perhaps EarthCube will support some effort like this. [08:25] AndreaWesterinen: Regarding my comment at 11:21 ... And, you need to be prepared to evolve the ontology and not religiously stick to the original definition. [08:26] TerryLongstreth: Matthew- "authoritative source is not necessarily an authority" puts a bigger burden on curation [08:26] Gary Berg-Cross: @SteveRay [11:20] I take the distinction to be that one is for a conceptual model and other is a representation is some logical formalism like OWL or CL for computation. [08:28] AndreaWesterinen: Regarding slide 17, "publishing" as is done by BBC, does indeed link concepts semantically, beyond just vocabularies. [08:28] Krzysztof Janowicz: See also Kuhn's paper on this topic: http://iospress.metapress.com/content/96636260086n02m6/fulltext.pdf [08:28] anonymous morphed into QuentinReu [08:29] QuentinReu: I disagree that publishing requires little semantics. More specifically, a lot of queries are not always determined during solution development [08:31] QuentinReu: Furthermore, the publishing industry is moving from just making documents available to representing nowledge within the content being published [08:31] QuentinReu morphed into QuentinReul [08:32] Krzysztof Janowicz: Slide 19 'rand' --> 'range' [08:36] Gary Berg-Cross: Note how much of the discussion in the Bottleneck Track is about Reuse, which is part of the Track A topic. [08:36] MikeBennett: Re Property domain and range point and questions: the domain and range problem can happen both ways. As @Matthew says you may find properties who D and R are set to a level of specificity for the app (e.g. hasAdverseEffect may relate to drug effect not effects more generally). Conversely you may find ontologies where D and R are set to "Thing" so that there is little more to the semantics than a label. [08:37] anonymous morphed into Mohsen Doroodchi [08:39] SundayOjo: Given that ontology of an application domain derives from the inherent semantic content of the application domain, ontology reusability in a target domain is constrained by semantic compatibnility of the two domains. This should not be seeing as a barrier, rather a fait accompli. [08:40] Gary Berg-Cross: @MikeB These examples of broad Domain (Thing) vs. narrow (Drug effect) reflect top down vs. bottom up approaches and we need something in the middle. [08:40] AndreaWesterinen: Re: domain and range ... We will talk more about this in the Track A discussion this afternoon. [08:43] pfps: pfps: An upper ontology can't be just a bundle of pieces. There are very high coherence and coverage requirements on upper ontologies. [08:45] MikeBennett: "Semantics is orthogonal to expressiveness" - @MatthewWest [08:46] MikeBennett: @Gary quite. The thing in the middle requires thinking about the actual meaning of the term. The top down and bottom up approaches are both workable technical approaches that work in a one-off application and don't require the developer to step into the unfamiliar territory of the philosophy of meaning. [08:46] SundayOjo: Semantic expressiveness of an ontology language should be measured in terms of the semantic gap between the specification model and the actual application semantics. [08:47] Krzysztof Janowicz: pfps, I would disagree. One example would be DOLCE versus the design patterns developed based on it. [08:50] Gary Berg-Cross: @pfps: A useful upper ontology might be aligned modular pieces. [08:51] PeterYim: == ChristophLange presenting the Ta=rack-B report ... [08:53] MatthewWest: @SteveRay: It just means the contributions came from different sources, not that something different is meant. [08:54] PeterYim: @pfps, would you be kind enough to morph into your real name please (use "Settings" button) at top center of the chat-window [08:55] MatthewWest: @AndreaWesterinen: That sounds a good approach. [08:55] SteveRay: @Matthew: Thanks. I'm also interpreting your distinction to be between the conceptual, logical and physical models, yes? [08:55] MatthewWest: @SteveRay: Yes. [08:57] MatthewWest: @TerryLongstreth: Yes, curation/provenance is in my view one of the big issues when the authoritative source is not an authority. [08:59] AndreaWesterinen: @MatthewWest and @TerryLongstreth [11:57] Curation and provenance is valuable regardless of the source - especially if the content is reusable across domains. The reuse could extend beyond the source. [08:59] ToddSchneider: QuentinReul, your problem space suggests a need for an architecture of ontologies to allow the use of the appropriate domain ontology to be used when there is a need to delve into the content of a publication. [09:00] ChristiKapp: I can only hear static [09:02] MatthewWest: @pfps: I agree an upper ontology needs coherence, none-the-less, it will consist of a number of discrete theories, each of which could be used in some other upper ontology. [09:03] Mark Underwood: @SundayOjo: Not a disagreement, but one um excuse for lack of reuse is the absence of ontology constructs in prevailing SDLC teaching & implementations [09:04] ToddSchneider: What would web-scale ontological commitments represent? [09:06] AmandaVizedom: I have to contest some of the statements on slides 9 & 10 [09:06] MatthewWest: @Andrea: Yes, curation and provenance in my view are prerequisites for reuse. [09:07] SundayOjo: Any assumption of universalism of world knowledge in formalisms is a simplistic assumption which is unrealistic. [09:08] pfps: I don't understand why OWL assumes universal knowledge. [09:08] AmandaVizedom: re: slide 9: As I understand it, from Chris Welty's talk and from another by Eric Moore, part of the Watson NLP team, Watson makes very heavy use of ontologies. Not by "turning big data into ontology" or to do heavy logical reasoning. Rather, in the NLP and disambiguation of word, phrase, statement, and data meanings. [09:09] pfps morphed into Peter F. Patel-Schneider [09:10] AmandaVizedom: re: slide 10: It is simply not true that traditional ontology (languages or content) assume universal knowledge. There is a school of ontology methodology that assumes this, but it is by no means widely or traditionally accepted within the ontology community. Partial and contextual knowledge are more commonly assumed. [09:10] PeterYim: @Peter F. Patel-Schneider, thank you, glad you are here! [09:12] Mark Underwood: @AmandaVizedom Agree - this was the description provided by a Watson team member at a NY Meetup I attended [09:13] Peter F. Patel-Schneider: I am puzzled as to how RDF can be considered to be an ontology langauge. RDFS, maybe, but RDF, no. [09:15] AmandaVizedom: @MarkUnderwood Indeed, the presentation by Eric Moore that I reference was also at a Meetup, in Durham NC (the Triangle SemWeb/AI/Machine Learning meetup). We discussed the point extensively. My understanding is that ontologies are used quite heavily in making meaning out of unstructured resources, especially. Quite similarly to approaches I am familiar with in Semantic Search & Retrieval over the web. [09:18] AmandaVizedom: (continuing my [12:15]) ... and this is quite compatible with what ChrisWelty said in his presentation to the Ontology Summit. He did not say that ontologies weren't used. He *did* say that they were used differently from how traditional research/AI ontologist think of using them. That matches with my understanding of Eric Moore's comments and my Semantic Search experience, in which the ontological relationships are used and matter very much to the text understanding, but there is not heavy duty logical reasoning beyond that. [09:19] Gary Berg-Cross: @Peter F. Patel-Schneider [12:13) I agree that RDF lacks real semantics to be called an ontology language. [09:20] QuentinReul: Where can information about "Linked Open Terms"? [09:20] Peter F. Patel-Schneider: @Gary It's not that RDF doesn't have semantics - it does - it's that RDF doesn't have any ontology facilities - no subclass, no domain/range, ... [09:21] Peter F. Patel-Schneider: ... these are, of course, all in RDFS, but RDFS is a very limited ontology language [09:22] MichaelGruninger: @Peter F. Patel-Schneider[12:08]. The phrase "universal knowledge" came from Alan Rector's presentation, and I think that it refers more to the idea that classical logics don't address typicality and defaults [09:30] Mark Underwood: @AmandaVizedom - good characterization. There is something to be learned from the reported difficulties IBM had in moving to health care from the core question-asking roots. It was discussed in the Watson Meetup but oh so lightly. [09:30] anonymous morphed into hypergrove [09:31] AmandaVizedom: I think that we need to distinguish between the richness, sophistication, and "weight" of an ontology and the kinds and weights of reasoning used over them. You can use combinations of ontological, logical, statistical reasoning and processing over data and metadata, all of which can be part of how the ontology is used. This aspect of how ontologies play in Watson may be a good indicator and source of lessons for us wrt how ontologies can or should be used in big data in general. [09:32] Peter F. Patel-Schneider: Michael The phrase is "traditional ontology languages assume universal knowledge." If "universal knowledge" means no typicality or defaults, then there has to be a lot more qualification, as many frame and rule systems also lack typicality and defaults. [09:33] Peter F. Patel-Schneider: I guess that it's lunch time. [09:33] ChristophLange: I just got disconnected from the conference (using Skype) [09:33] ChristiKapp: I am disconnected too [09:33] MikeBennett: In the roomwe have music after a message about the moderator having disconnected. [09:33] MichaelGruninger: @Peter F. Patel-Schneider: Agreed. We should pester Alan Rector about this some more :-) [09:33] ChristophLange: in response to the last comment (which probably continued beyond the disconnect): ... [09:34] David Blevins: stepping out for lunch. Resuming at 2 [09:34] NSF-venue: I think the remote session timed out after 4 hours. We will reconnect after lunch. [09:34] NSF-venue: Return at 2pm Eastern US time. [09:34] ChristophLange: You (Steven?) appropriately characterised the problems with governments etc. pushing towards open data, but not saying much about their semantics, and of the lack of best practices [09:35] ChristophLange: W.r.t. "minimal commitments" I think we see a number of helpful approaches already [09:36] ChristophLange: 1. There is the RDF/Linked-Data approach, whose logic has little expressiveness and in that sense doesn't enforce too many commitments, and there are lots of best practices around this ... [09:36] ChristophLange: ... including a new W3C working group "Data on the Web Best Practices": https://www.w3.org/2013/dwbp/wiki/Main_Page [09:37] ChristophLange: 2. OntoIOp is still an emerging standard and, granted, it's not yet that widely adopted that one could speak of "best practices", but at least it allows you to develop small ontology modules, each of them in the ontology language that's most appropriate. [09:38] ChristophLange: Some more answers/comments to points from the chat during my talk (while those present in person are hopefully enjoying their lunch!) ... [09:40] anonymous morphed into DimitriDarras [09:42] ChristophLange: @ToddSchneider [12:04]: Web-scale ontological commitments could mean to aim at a big unified ontology of everything that's out there on the Web. Like an upper ontology, of which all big data "out there" are valid instances. This is something that seems unrealistic to me. E.g., in the linked data world, DBpedia (distilled from Wikipedia) does describe all domains, and it is a preferred target of being linked to from other ontologies/data-sets, but the descriptions and the links are at a rather informal level, often not doing much more than disambiguation. If DBpedia's descriptions were formally stronger, a lot of links to DBpedia from other datasets would no longer hold. [09:44] ChristophLange: Re "universal knowledge", I confirm that MichaelGruninger [12:22] explained the meaning of the term that was intended here (rooting in AlanRector's 2014-01-30 presentation) [09:45] ChristophLange: @Peter F. Patel-Schneider [12:13]: RDF is certainly a very weak ontology language, but it is a formal language with a logic-based semantics, which is used to express knowledge (or, more modestly, to structure data), so why not call it "ontology language" in a more general sense? [09:47] ChristophLange: @QuentinReul [12:20]: For Linked Open Terms (LOT) please see http://lot.linkeddata.es/ [11:02] anonymous morphed into JimSolderitsch [11:06] QuentinReul: Is there a hashtag for the meeting? [11:06] QuentinReul: @ChristophLange thanks for the link [11:06] CoreyLeong: #ontosum [11:10] SimonSpero1 morphed into SimonSpero [11:14] PeterYim1: == PhilipBourne keynote talk on Data at NIH ... (started for a few minutes now) [11:18] Gary Berg-Cross: I think that Bourne's slides are at http://www.slideshare.net/pebourne [11:19] anonymous2 morphed into Stephane Fellah [11:37] Gary Berg-Cross: Bourne is addressing the data sustainability challenge noted earlier at [10:09] Gary Berg-Cross: Long-term preservation of data.. [11:38] Gary Berg-Cross: I believe that other countries like Australia do support data preservation on more of a national basis. [11:46] Mark Underwood: Premise: Use of Complex Event Processing on Big Data is inversely proportional to privacy [11:50] David Blevins: there are some systems which allow data obfuscation rules to be defined through existential quantification [11:52] SteveRay: Regarding the privacy issue, see the NYTimes article: http://www.nytimes.com/2014/04/27/technology/a-student-data-collector-drops-out.html [11:55] Stephane Fellah: http://www.myexperiment.org/ seems to address some of the aspects outline in the presentation [11:58] Mark Underwood: Re: Communities of Interest - In the Elsevier Innovation Explorers grp I participate in, there is a natural conflict of interest between publishers, institutions, professional associations and individual researchers. Each has a stake in the data and in owning it somehow. Example is institutions standing up their own ResearchGates. Institution-based systems may win out if you are tenured [11:58] List of members: AndreaWesterinen, AnneThessen, BobbinTeegarden, BobbinTeegarden1, ChristiKapp, Christof Hasse, ChristophLange2, CoreyLeong, DanCerys, DanCorbett, David Blevins, DeborahLNichols, DimitriDarras, ElizabethFlorescu, FrankOlken, FrankOlken1, Gary Berg-Cross, hypergrove, Jesse Wang, JimSolderitsch, JulioRoa, KenBaclawski, LeoObrst, Mark Underwood, MikeBennett2, MikeRiben, MosesLesibaGadebe, NaicongLi, NSF-venue, PeterYim, PeterYim1, QuentinReul, SimonSpero, Stephane Fellah, SteveRay, SundayOjo, Todd Pehle, ToddSchneider, TorstenHahmann, vnc2 [11:59] QuentinReul: Has the term of "data revolution" just been coined? [12:02] DanCorbett: The term "Information Revolution" has certainly been around for a long time (decades). [12:03] MikeBennett2: Semantic revolution, on the other hand... [12:04] Mark Underwood: The same parents who complained about InBloom are often perfectly happy to hand over data to test prep centers, Common App, the College Board - often coupled with 1040 data elements. [12:09] QuentinReul: Would imagine that initiative, such as DOAP ( https://github.com/edumbill/doap/wiki ) would cover some of the needs for the NIH repository [12:14] ChristiKapp: Track A - Slide 7 [12:26] CoreyLeong: on the flip side, identifying antipatterns with ontologies may help also [12:30] Mark Underwood: Sowa paper: Historical perspectives on problems of knowledge sharing tp://bit.ly/1itlNLL [12:31] Mark Underwood: http://bit.ly/1itlNLL [12:44] anonymous morphed into JimDisbrow [12:54] NSF-venue: Next session will begin at 4pm Eastern US [12:58] KenBaclawski: I am on the conference line and I can hear the meeting, but it doesn't appear that you can hear me. Can you check this? I did unmute myself. I will be speaking remotely at 4pm. [13:00] AnneThessen: I cannot dial in to the conference line [13:01] AnneThessen: I'm here now [13:03] NSF-venue: Just about ready to start [13:04] NSF-venue: @Ken: Can you try speaking again? [13:04] NSF-venue: Yes [13:06] NSF-venue: == Report from Track D, KenBaclawski == [13:21] anonymous morphed into JimDisbrow [13:22] ToddSchneider: Question about provenance of data: to provide an adequate provenance of where data came from, or how it was created, is there a need to make use of knowledge/ontologies from domains other than those of the initial intended uses of the data? [13:31] Gary Berg-Cross: Of relevance to this Track topic will be a proposed panel at the NSF All-Hands EarthCube meeting June 24-26 Conveners: Gary Berg-Cross, Pascal Hitzler, Kerstin Lehnert, Peter Wiebe Panelists: Anthony Aufdenkampe, Tim Finin, Krzysztof Janowicz, Scott Peckham Our preliminary title is "Addressing heterogeneity of information representations in EarthCube" [13:41] Mark Underwood: Sad fact about provenance concerns: in the headlong plunge to push data into Hadoop & the like, provenance takes a back seat. [13:42] MikeBennett1: @Mark so how does one distinguish data from truth? [13:42] Mark Underwood: @MikeB: Yeah. [13:44] NSF-venue: == Report from Track G, AmandaVizedom and OliverKutz == [13:45] Mark Underwood: OTH a great presentation by Brocade at the San Diego BDWG / ISO JTC meeting argued that a key app for Big Data IS provenance - that an entire network stack dump (in the SDN sense) is needed to ensure provenance [13:47] NSF-venue: == Report from Track R, AmandaVizedom, MarcelaVegetti == [13:55] Mark Underwood: Putting the slide decks on Slideshare (LinkedIn) could get some add'l views. That team is very SEO savvy [13:59] David Blevins: @KenBaclawski forgot to ask a question during your Track-D report. Do you think the lack of interest in ontology semantics, and general lack of understanding related to ontologies, is a matter of a lack of knowledge related to the technologies, or a perception that the technology is not useful? (Perhaps a bit of both?) [14:00] PeterYim: == Report from Track R: Outreach - Ontology Summit semantic (psmw) wiki-site KenBaclawski, TejasParikh, SimonSpero, PeterYim [6 min] (PeterYim presenting) No slides, I will go through the following pages on the OntologPSMW in this presentation: - http://ontolog-02.cim3.net/wiki/OntologySummit2014_symposium - http://ontolog-02.cim3.net/wiki/OntologySummit2012 - http://ontolog-02.cim3.net/wiki/OntologySummit2013 - http://ontolog-02.cim3.net/wiki/Category:OntologySummit2013_Survey - http://ontolog-02.cim3.net/wiki/OntologySummit2013_SurveySummary - http://ontolog-02.cim3.net/wiki/SandBox#Tests - http://ontolog-02.cim3.net/wiki/WikiHomePage - http://ontolog-02.cim3.net/wiki/OOR [14:08] NSF-venue: == Report from Track M: ChristiKapp == [14:17] TimFinin: People are trying to build tools to automatically capture provenance, see PROB: A tool for Tracking Provenance and Reproducibility of Big Data Experiments, http://bit.ly/1lqBlzJ [14:19] KenBaclawski: @DavidBlevins Here are my own personal observations. The first problem is the "Not in my bailiwick problem". Data scientists see semantics as the problem of the producers of the data. The producers of the data have no background in data semantics and would probably be very surprised if anyone suggested that they were responsible for data semantics. Data scientists do seem to know that data semantics and ontologies exist even if they may not know much about them. There is also the problem that there is a perception that ontologies are only used for helping with search, even to the point of equating ontologies and search. [14:20] NSF-venue: == Remarks from PeterYim: The Ontology Summit Operation in Perspective == [14:34] David Blevins: @KenBaclawski: thanks for the feedback. Any thoughts on solutions to these issues? Do you think that the lack of visible semantic web and ontological solutions may be creating the issue of perceived utility? I occasionally have issues demonstrating the merits of ontologies to data scientists, and the lack of highly visible examples creates quite a bit of work in convincing them. [14:34] David Blevins: utility outside of search* [14:37] Gary Berg-Cross: Ontologies are not very visible in the same way infrastructure is not easily understood by some. [14:38] SimonSpero: TinFinin: http://www.eecs.harvard.edu/syrah/pass/ (provenance aware storage systems) [14:40] SimonSpero: ( also http://www.eecs.harvard.edu/margo/papers/ ) [14:41] JimSolderitsch: Verticalnet -- that's a name from my past! [14:43] MarcelaVegetti: Thank you Peter! [14:44] ChristophLange2: Thanks, Peter, and congratulations -- enjoy the dinner! [14:45] NSF-venue: We are adjourned for today. Thanks to all for attending today. [14:48] KenBaclawski: @DavidBlevins: I don't think that there will be a single solution in all cases. The use cases I presented in my talk are examples of solutions. The hope is that we can harvest examples of successful solutions in the Ontolog wiki and that we can use them as a resource to demonstrate the merits of ontologies as well as use them as techniques for solving these issues. [14:48] ChristiKapp1: Congratulations Peter ! [15:14] List of attendees: AmandaVizedom, AndreaWesterinen, AnneThessen, BobbinTeegarden, BobbinTeegarden1, BobbinTeegarden10, BobbinTeegarden11, BobbinTeegarden12, BobbinTeegarden13, BobbinTeegarden14, BobbinTeegarden15, BobbinTeegarden2, BobbinTeegarden3, BobbinTeegarden4, BobbinTeegarden5, BobbinTeegarden6, BobbinTeegarden7, BobbinTeegarden8, BobbinTeegarden9, ChristiKapp, ChristiKapp1, ChristiKapp2, Christof Hasse, ChristophLange, ChristophLange1, ChristophLange2, CoreyLeong, DanCerys, DanCorbett, David Blevins, DeborahLNichols, DimitriDarras, ElizabethFlorescu, FrankOlken, FrankOlken1, FrankOlken2, FrankOlken3, FrankOlken4, FrankOlken5, Gary Berg-Cross, Jesse Wang, Jesse Wang1, Jesse Wang2, JimDisbrow, JimSolderitsch, JulioRoa, KenBaclawski, Krzysztof Janowicz, LAPritchard, LeoObrst, LeoObrst1, LeoObrst2, LeoObrst3, MarcelaVegetti, Mark Underwood, Mark Underwood1, Mark Underwood2, Mark Underwood3, Mark Underwood4, MatthewWest, MichaelGruninger, MikeBennett, MikeBe nnett1, MikeBennett2, MikeBennett3, MikeBennett4, MikeRiben, Mohsen Doroodchi, MosesLesibaGadebe, MosesLesibaGadebe1, NSF-venue, NaicongLi, NancyWiegand, PavithraKenjige, Peter F. Patel-Schneider, PeterYim, PeterYim1, PeterYim2, QuentinReu, QuentinReul, SimonSpero, SimonSpero1, Stephane Fellah, SteveRay, SteveRay1, SundayOjo, TerryLongstreth, TerryLongstreth1, TimFinin, TimFinin1, Todd Pehle, ToddSchneider, TorstenHahmann, TorstenHahmann1, TorstenHahmann2, anonymous, anonymous1, anonymous2, anonymous3, coreyleong, hypergrove, leo, mosesg, pfps, vnc2 ------