ppy/symposium-day-2_chat-transcript_unedited_20120413a.txt ------------- Chat transcript from room: ontologysummit2012 2012-04-13 GMT-08:00 ------------- [06:07] PeterYim: Welcome to the = OntologySummit2012 Symposium - Day 2 - Fri 2012-04-13 = Summit Theme: OntologySummit2012: "Ontology for Big Systems" Ontology Summit 2012: General Co-chairs: Dr. LeoObrst & Dr. NicolaGuarino Ontology Summit 2012: Symposium Co-chairs: Dr. RamSriram & Professor MichaelGruninger Session page: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2012_Symposium Mute control: *7 to un-mute ... *6 to mute Can't find Skype Dial pad? ... it's under the "Call" dropdown menu as "Show Dial pad" . == Proceedings: == . [05:49] anonymous morphed into Amartinez [06:17] anonymous morphed into Hans Polzer [06:22] SteveRay1 morphed into SteveRay [06:26] anonymous morphed into DeborahMacPherson [06:48] BrandNiemann: See http://semanticommunity.info/A_NITRD_Dashboard/Research.gov and http://semanticommunity.info/A_NITRD_Dashboard/Semantic_Medline [06:57] FrankOlken: I have read a few dozen data management plans for NSF while reviewing proposals. They are generally quite mediocre. No more than [06:57] RaviSharma: Line- your comment pertains only to science data or all big data [06:58] FrankOlken: No more than 10 percent of the data management plans discuss metadata - less than 5 percent mention terminology or ontologies. Most of the proposals I read for NSF are computer science data management research proposals. [06:58] AmandaVizedom: @Frank - ouch! [07:02] Rex Brooks: @Frank: Do you think this is due to lack of familiarity with ontology/semantics or a dismissal of anything that hasn't proven its acceptability to the readers at NSF? [07:04] Rex Brooks: BTW this doesn't imply any prejudice among NSF readers, just among grant writers. [07:05] KenAllgood: @Frank - That's exactly one of the missing pieces that we've identified as we build the data management strategy for the DoD/VA integrated health record.. In looking at existing DMSs and DMPs within the federal space there is little mention or consideration of the need for metadata definition/management, and even less mention of taxonomy/ontology structural definition/management. They tend to be either in the stratosphere or focused only upon logical and physical layers.. [07:08] RaviSharma: Geroge - Semantic medline is excellent also for rural and remote practitioners, for DOD field situations for Haiti type preparatory early teams etc. [07:09] BrandNiemann: http://semanticommunity.info/A_Quint-Cross_Information_Sharing_and_Integration/2010_Haiti_Earthquake_Metamodel [07:11] RaviSharma: George, Leo, others - Agencies as they create operational budgets ought to be guided by OMB etc for at least annotating metadata more exhaustively for big data and standards now more semanttic oriented need to be specified so the we can access archived data later as well as absract only the useful info for a given purpose through ontology reasoning filters etc. [07:11] SimonSpero: @Frank - I have been thinking about this recently after reading some of the questions asked on some of the digital curation mailing lists by people wondering how to do a proper DMP. I'd love to do a study of grant applications and dmp, and see if there's any features that are predictive of ability to survive various plausible scenarios, and to see if there are any features that are predictive of grant outcome (hard to do without having access to identity of applicants and their base-line success rate). Darned natural experiments [07:12] AmandaVizedom: @Frank, Rex - I have no inside info, but I had the impression last year that NSF had *some* ideas as to how data management plans might work, but is more heavily looking for the most promising approaches to emerge from the research communities and in the proposals. It's also been fairly apparent that most researcher, even in neighboring technical communities to this forum, have little idea how to approach this and are hoping/waiting for guidance. Or for someone's approach to succeed and be reusable. Consider how little semantic technologists in one application area tend to know about other application areas; it's not so surprising that even the computer science data management research proposals are not showing evidence of contribution from people with sufficient knowledge of semantic metadata to propose an DMP along these lines. [Still: ouch! Such a promising direction.] [07:14] SimonSpero: The dissertation I was mentioned is by Dr. Hollie White (who is now at Duke Law) http://sites.duke.edu/holliewhite/dissertation-research/ [07:15] RaviSharma: Brandt - you are unbelievably great -thanks for a great link, that is semantic communication instantly. regards. [07:16] SimonSpero: [Hollie is looking to see how the results apply to the gap between law libraries/academic lawyers and the information needs of practitioners]. Electronic Discovery (ESI) also rears its head. [07:18] SimonSpero: I don't like Medline being a "killer app" [07:18] RaviSharma: Line - funding mendated by these requirements could be helpful to such groups if they want research to progress perhaps for cross disciplinary learning. [07:19] LeoObrst: @Frank: I don't know about NSF, but I have understood other funding agencies to frown on funding any ontology development effort, choosing instead to place their money on computational research (that possibly use ontologies). [07:19] MikeBennett: I wonder if there's a human aspect to the lack of semantics in data management project proposals: what it takes to be any good with semantics is a very different skill-set to what it takes to be god with data. Data management is a business problem not a technical one, and once it is elevated to the business and out of IT, then at least you would expect to see some appreciation of semantics. [07:22] LeoObrst: In addition, many database folks think they are already doing semantics (though they are not). This stance is often misleadingly underscored by ontology-savvy folks who tend to conflate the use of logic with the use of ontology. [07:22] MikeBennett: @Leo +1 and then some! [07:34] KenAllgood: @Leo I see this everyday, and the misunderstanding has driven a stake through the hearts of many potentially promising efforts. [07:35] BrandNiemann: Thanks, Ravi, and Best Regards [07:35] PeterYim: ALL: note that the demo from DoD will be delivered through the voice conference line with shared screen support using the GotoMeeting service (courtesy of the EDMCouncil / MikeBennnett - Thank you for the support) ... please try to get connected a few minutes before the demo starts - see: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2012_Symposium#nid39VO [07:41] TerryLongstreth: @Leo: can you expand on the differences between what DB folks do and what they think they do, and the use of logic vs. the use of ontology? [07:43] AmandaVizedom: This discussion has returned several times to emphasis on distinguishing semantic metadata and semanticizing primary data. This seems to me quite similar to a common case in federated data and semantic interopability. In both cases, the immediate need and clear payoff are often in semantic metadata, not changing raw science data or legacy enterprise data, for example. There is a lot of benefit for ontological wrappers that provide information about how to interpret the raw or legacy data, specifying what it means. This enables accurate use/reuse of the data and opens it up for semantic discovery and query. [07:49] BrandNiemann: I call it semantic enhancement! [07:49] FrankOlken: @LeoObrst NSF has supported some ontology development efforts. Generally ontology development for specific disciplines is considered the responsibility of the particular subject area program (e.g, biology or earth sciences rather than computer science). NSF has spent very little on development of general ontologies, more commonly on development of semantic web tools / systems, some work on logic programming, inference engines, logics (various description logics), etc. [07:51] FrankOlken: @Leo For example: NSF supports Michael Kifer (logic programming), Pascal Hitzler (description logics, inference algorithms), Deborah McGuiness (semantic tools for Virtual Observatories), ... [08:06] FrankOlken: Yes, I can see the slides for the demo of the dashboard. [08:10] PeterYim: In case anyone have trouble with GotoMeeting, the shared screen is on the VNC screen too (now) [08:12] Rex Brooks: This DoD app is exactly what I've been working toward for the last few years, so I'm gratified to see that I am not far off base. [08:13] Rex Brooks: Especially since emergency management and the military have a large overlap that I have to take into account. [08:14] LeoObrst: @Frank: sure, these are computational uses of ontologies, semantics, but the ontology development per se is generally limited, focused on automated reasoning. [08:19] FrankOlken: @Leo Yes, NSF CISE (Computer & Info Science and Engineering) considers content development (ontologies, databases, etc.) outside of its scope - such proposals are [08:19] FrankOlken: such proposals are referred to subject area directorates (bio, geo, etc.). [08:20] FrankOlken: Conceivably, generic ontologies (time, space, processes) might be supported by CISE IIS. [08:23] FrankOlken: Leo is speaking (I think). [08:23] FrankOlken: Who is speaking? [08:24] anonymous morphed into EvanWallace [08:24] FrankOlken: Is Ed Barkmeyer speaking? [08:25] GaryBergCross: Ed Barkemeyer is speaking. [08:25] LeoObrst: Yes, that was me and Ed talking. [08:28] FrankOlken: Amanda Vizedom is speaking. [08:29] GaryBergCross: The correct spelling is Barkmeyer, I believe. [08:34] LeoObrst: @Terry: Yes, db folks typically go through a 3-schema development: conceptual, logical, physical, with left-to-right refinement. The conceptual schema is as close to a semantic model as occurs, but it is typically "local semantics" which contains local constructs for the company, mission, system, database (vs. "real world" semantics of ontologies). And then no linkage is kept from the conceptual to the logical to the physical models, but instead the first 2 models (often like UML models in object systems after programming) are thrown away, used only for documentation (along with the data dictionary) and never maintained. Instead thereafter the physical schema is modified, and so the whole process is not "model-driven". So the result is a "structural model". [08:38] anonymous morphed into line pouchard [08:39] line pouchard morphed into LinePouchard [08:40] AliHashemi: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2012_Symposium [08:46] LeoObrst: @Terry: more to your second question: because rdbs use the relational calculus/SQL as the query language, often some ontology-savvy folks will conflate the use of that logic with the use of ontology/semantics. But the non-logical terms used in the query language may or may not actually be part of a coherent ontology/ontologies, but instead syntactic/structural terms from the physical schema, i.e., more like a vocabulary. The relational calculus is actually a subset of first-order logic, but is clearly related. In a similar way, logic programming using Horn logic is a syntactically restricted FOL with some operational semantics (cut, order restrictions, etc.), though more expressive than the rel calc. However, it is also the case that simply using Prolog does not mean you are using ontologies. [08:51] LeoObrst: @Doug: finally, DB folks may use ontological analysis when they develop their conceptual schema, and the best ones probably do, but it really is a truncated analysis, since they know their output is not an ontology but instead a physical schema. [08:55] DeborahMacPherson: Great Gary! There has been a lot of progress with SOCoP [08:55] GaryBergCross: Thanks!...I hope that we can keep it going. [08:55] MikeBennett: @Gary you wanted to know if we have stuff. We have stuff. [08:57] LeoObrst: Actually was to @Terry! [08:58] GaryBergCross: @Mike Yea!! Let's talk.... [09:03] DeborahMacPherson: Does OOR require Flash? [09:04] PeterYim: @Deborah - No, OOR does not need flash [09:18] DeborahMacPherson: It keeps asking me to download a plug in - then can't find it - what might it be? [09:21] SimonSpero1: Links & Resources:PRIKL - http://sourceforge.net/projects/prikl/ : HETS toolkit is http://www.informatik.uni-bremen.de/agbkb/forschung/formal_methods/CoFI/hets/index_e.htm [09:21] RaviSharma: Mike Grunninger- I am very happy at the variety of ontologies that you have now enabled on OOR using Calore etc. Thanks for your efforts. [09:22] SimonSpero1: PRIKL requires java and maven : HETS needs the Glasgow Haskell Compiler [09:24] PeterYim: @Deborah - is this related to visualization ... they may have different plugins that support visualization (say, similar to Protege) [09:24] DeborahMacPherson: maybe [09:25] AmandaVizedom: @MichaelGruninger - very exciting to see COLORE, in particular the verification tools and results. I hope that integration of evaluative metadata into ontology repositories, along with tools to generate that metadata, continues to evolve. Even many ontology creation tools, in which one might most expect such tools for development-time validation, tend to have very little of use. Even where they have more, however, having such capabilities in the repository amounts to the availability of search-time, independent evaluative services -- much more useful to those searching for ontologies to use. [09:29] ToddSchneider: What is 'new science'? Will conflict with the current academic stove pipes? [09:35] GaryBergCross: I think that the new science idea is data-centric and includes analytics and discovering patterns in the data. [09:39] SimonSpero1: "New science" is an application of "New Math", http://www.youtube.com/watch?v=8wHDn8LDks8 [09:40] TerryLongstreth: @Leo - Thanks. [09:41] SimonSpero1: http://www.nsf.gov/funding/pgm_summ.jsp?pims_id=504767 [09:42] RaviSharma: Dr. George Strawn and Ms. Wendy Igen - great effort looking forward to discussion on Sytem architecture nexus [09:43] FrankOlken: The NSF solicitation mentions $25M dollars. [09:44] FrankOlken: Most of the money for the NSF-NIH solicitation (about 90%) is coming from NSF. [09:44] FrankOlken: I suspect that the bulk of the funds for the NSF-NIH solicitation will come from NSF-CISE - there are many other NSF directorates participating, [09:45] FrankOlken: but I am skeptical how much cash they are putting up. [09:45] GaryBergCross: The 4th Paradigm book is readable online at http://research.microsoft.com/en-us/collaboration/fourthparadigm/4th_paradigm_book_complete_lr.pdf [09:48] RaviSharma: George - Perhaps value comes from combining computation and big data management to provide new scientific opportunities that individual scientists working with limited access to data may not be able to achieve. [09:49] ToddSchneider: Gary, Thank you. [09:49] FrankOlken: No, I do not believe that IMLS is involved. They are not considered science. [09:49] GaryBergCross: The diff of oil import by year reminds me of a conceptual analysis problem a bit more than an ontological analysis. [09:51] FrankOlken: Actually, the CISE contact on the NSF-NIH solicitation has some background in semantic web technology, Vasant Hanovar, but he is primarily a data mining person. [09:51] SimonSpero1: IMLS is http://www.imls.gov/ [09:52] SimonSpero1: They were involved with NAS BRDI http://sites.nationalacademies.org/PGA/brdi/index.htm , etc. [09:53] FrankOlken: NSF does not generally fund researchers outside the U.S. (except occasionally as consultants to U.S. based projects). [09:53] SimonSpero1: Much smaller funding agency than NSF and NIH [09:53] RaviSharma: Wendy - high school and college students - at least some will be our teachers - they will also solve the problems that we have created for the next generation such as transportation, environmental challanges. [09:54] FrankOlken: It is not that IMLS is small - it is considered to be part of the humanities community not science. [09:54] AliHashemi: ontolog.cim3.net/OntologySummit/2012/communique.html [09:55] AliHashemi: the above link ^^ is the final version of the communique. [09:55] GaryBergCross: I think that Wendy's slide 6 gives a reasonble view of New Science - A future in which the ability to analyze and extract information from large, diverse, and disparate data sets accelerates the process of scientific discovery and innovation; promotes new economic growth; and leads to new fields of research and new areas of inquiry that would otherwise be impossible. impossible. [09:55] AliHashemi: or if you want to view it directly in google: https://docs.google.com/document/d/1OG_iNRROkfh2T76Ri0SrNzwLwVKGKo4kQOWwBKxHjy8/edit [09:58] MikeBennett: The GoToWebinar session has now been closed - after lunch we'll use the normal mechanisms for these things. [10:02] PaulHunter asks: will i Magic 8-Ball says: As I see it, yes [10:53] Rex Brooks: The last sentence is cute. Do we want to be cute? It does make a point about semantics, but... [10:55] Rex Brooks: I will still be happy to endorse/sign it. [10:58] anonymous morphed into TomTinsley [11:03] Rex Brooks: I'm sure it was meant to be humorous. [11:04] FrankOlken: I seem not be able to be heard. I have problems with the last paragraph in Section 4.3 advocating hiding the ontologies and using SKOS. [11:11] FrankOlken: I would suggest we urge the use of controlled natural language instead of raw logic. [11:13] FrankOlken: Yes, I wouldlike to see the sentence changed. [11:13] FrankOlken: or possibly stricken. [11:14] FrankOlken: Accurate but more vernacular presentation. [11:15] MikeBennett: Accurate but simple view using diagrams and controlled natural language, as appropriate to their context. [11:15] MikeBennett: (for example} [11:15] FrankOlken: Mike's version would be okay for me. [11:16] MikeBennett: Did you do a #7 or whatever Frank? [11:16] FrankOlken: @Mike Yes. I am unmuted. [11:16] FrankOlken: @Mike unmuting is *7 [11:17] MikeBennett: Oops. [11:18] Rex Brooks: I believe Leo made the best suggestion. Remove the several sentences about broad agreement and just say that the presentation should be relevant to the users' context. [11:18] Rex Brooks: I love bananas. [11:18] Rex Brooks: I bet fruits flies do, too. [11:19] CoryCasanave: @Rex, +1 [11:20] FrankOlken: Yes. [11:21] FrankOlken: With the change to Section 4.3 I can endorse the communique. [11:22] MikeBennett: I cannot see what the changes to 4.3 were, it still looks the same, but if the changes we discussed (per Frank, Leo, Rex etc. above) will be in it I'm certainly happy with it. [11:23] anonymous morphed into SteveRay [11:23] MikeBennett: Spoke too soon, now it looks perfect. [11:27] ToddSchneider: A Facebook page for the Communique? [11:28] FrankOlken: I am reasonably happy with Sect. 4.3 now. [11:29] FrankOlken: I have already sent in my endorsement to the ontology summit list. [11:29] Rex Brooks: I endorse it, and I will also write in to the summit list to make sure it is registered. [11:32] FrankOlken: section 4.4. is present in the Google Docs version [11:32] DougFoxvog: (and (similarTo (PassageOfFn Time) (FlightFn Arrow)) (likeToEat FruitFlies Banana)) [11:35] DougFoxvog: Or: (and (similarTo (PassageOfFn Time) (FlightFn Arrow)) (attractedTo FruitFlies Banana)) [11:38] Rex Brooks: We may be getting close to time to do some "Where We Are Now" in relation to previous summits. I am working on ontologies in standards, looking toward a more comprehensive domain ontology with sub ontologies for vocabularies and codelists for Emergency Management. Whether I will get this done by next year is uncertain, but I will eventually. [11:39] PeterYim: == Suggestions for future Ontology Summit themes ... (please enter below0 [11:39] Rex Brooks: We could look at educating ontologists, making the case, etc. [11:39] Rex Brooks: Or we could update any of our previous summits. [11:39] MichaelGruninger: Perhaps we can draft a document that proposes a new SIG on engineering and ontologies within IAOA (International Association of Ontology and its Applications) to be discussed at the FOIS conference in July. The conference will also have a special track on engineering and ontologies. [11:40] MikeBennett: Shared semantics and standards - I know we covered it previously but this is now reaching a critical mass of interest across a few industries including ours. [11:40] MikeBennett: Whoever said autogenerating ontologies from data: Noooo! [11:41] ToddSchneider: Suggestion for next year's summit, automatic ontology generation [11:41] AliHashemi: Proposal: How do ontologies fit into broader knowledge systems. How exactly are they components of bigger systems? [11:42] MikeBennett: "Ontologies for normal people" (Cory) +1 [11:43] BobbinTeegarden: Suggestion for next year's summit, also: executable ontologies and incorporation of process; visualization of ontologies [11:43] GaryBergCross: George suggested that a bottleneck was developing domain vocabularies and that NLP had advanced where we might be able to use such tools to start on a voc on the way to an ontology. Last year Doug Skuce talked about FactGuru which would do this type of thing. [11:44] Rex Brooks: Ontologies in Standards and as Standards would include Architecture, and like Geospatial ontologies, need to move toward a convergence on an upper ontology or ontologies. [11:46] CoryCasanave: Like @Bobbins suggestion also [11:46] GaryBergCross: Possible topic for next Summit - Nicola proposed something on the social-cutural area. His talk on Social Roles would be one part of such a discussion. This would be broadly relevant and has been explored in areas like Law and Land use, fiat boundaries, ownership etc. [11:47] SimonSpero: My suggestion was: "Managing Ontology Based Projects" [11:47] MikeBennett: Coulc ontologies in a socio-cultural context extend to socializing upper ontologies, visualizing ontologies, using standards and so on? [11:47] Hans Polzer: Regarding suggestions for future summits, consider following up on Amanda and Mike's talk on Ontology Quality Practices, especially the "fitness for use" quality attributes. Specifically, consider developing an ontology for representing the scope of applicability of ontologies (including itself!). Note that there are probably multiple such ontologies. Such an ontology would also be very useful in facilitating finding the right ontology in a repository and applying ontologies to problems, including the big data problem. [11:47] Rex Brooks: Deriving ontologies from social media would be interesting, but I doubt it is sufficient for a summit on its own. [11:47] DeborahMacPherson: Suggestion - ontology driven architecture and data integration [11:47] AliHashemi: Ravi: Idea is to develop ontologies that test for standards compliance or validation such as NIST reflectors? [11:48] MikeBennett: @Hans sort of a metaontology? Awesome! [11:48] AmandaVizedom: Comment: there many tools and research projects for generating "ontologies" from data, other sorts of models, and corpora. There is a big difference between creating such a thing and creating an ontology that captures the actual meaning of the data, e.g. Without good methods and tools for evaluating semantic fit (as in the current situation), this kind of generation is mostly a source of things in owl that aren't really ontologies, or things that are ontologies but aren't the needed ontology. [11:49] MikeBennett: @Amanda precisely - there are still enough people out there who think syntax endows semantics, that it would misleading to even explore this route just yet, I think. [11:49] Rex Brooks: @Amanda: Yup. Just looking at social media as a symptom of unexpressed, implicit ontologies. [11:49] MichaelGruninger: Ontologies for Semantic Integration. [11:50] GaryBergCross: @MikeB "socializing upper ontologies" might not be the term everyone would use, but there is work to add some of these social-cognitive primitive as equal to physical constructs like object. [11:51] AmandaVizedom: ... actually, that's true regardless of whether the generation is automatic or manual. Perhaps the more important point is that ways to create more ontologies faster aren't necessarily good without significant progress in evaluation. [11:51] Rex Brooks: Thanks All! [11:51] BobbinTeegarden: Excellent work, kudos to everyone! [11:52] SteveRay: Various comments made about integration, reference artifacts, challenges, Olympics, [11:52] GaryBergCross: One Description of Social Relations is at file:///c:/users/Gary/Documents/Conferences/FOIS/Descriptions%20of%20Social%20Relations.htm [11:53] GaryBergCross: I guess we get time off for productive behavior... [11:54] SimonSpero: Gary - that's a file: url .. [11:56] AliHashemi: Thank you all. Bye [11:57] PeterYim: -- session ended: 2:45pm EDT -- -------------