ppy/chat-transcript_unedited_20131024a.txt ------ Chat transcript from room: ontolog_20131024 2013-10-24 GMT-08:00 [PDT] ------ [9:28] PeterYim: Welcome to the = RulesReasoningLP: Mini-series Launch Event - Thu 2013-10-24 = Topic: RulesReasoningLP: Mini-series Launch Event - Survey and Introduction ) Session Co-chairs: Dr. LeoObrst (Ontolog; MITRE) & Dr. BenjaminGrosof (Coherent Knowledge Systems) Opening Remarks by Community & Technology Leaders and the Mini-series Co-champions: * Professor MichaelGruninger (IAOA; U of Toronto) * Professor MichaelKifer (SUNY, Stony Brook) * Dr. LeoraMorgenstern (SAIC) * Dr. VinayChaudhri (SRI) * Dr. HaroldBoley (RuleML; U of New Brunswick) * Dr. HensonGraves (Algos Associates; OMG) * Professor KenBaclawski (Northeastern U) * Dr. JohnSowa (VivoMind Research) * Mr. MikeDean (Raytheon-BBN) * Mr. PeterYim (Ontolog; CIM3) Survey and Introduction to Key Concepts and the Technology Landscape * Dr. LeoObrst (MITRE; Ontolog) - "Survey: Logic, Logic Programming, Ontology, Rules" * Dr. BenjaminGrosof (Benjamin Grosof & Associates) - "Survey of Knowledge Representations for Rules and Ontologies" Please refer to details on the session page at: http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2013_10_24 . == Proceedings: == . [9:26] anonymous morphed into BrandonWhitehead [9:29] anonymous morphed into Weihong Song [9:31] AmandaVizedom: Logistics note, confirming what Peter mentioned last week: you can skype-call to "joinconference" even if it shows as invisible. I've just gotten on the call that way. [9:31] anonymous morphed into ChristopherSpottiswoode [9:32] anonymous1 morphed into Adrian Giurca [9:32] anonymous morphed into NaicongLi [9:32] anonymous1 morphed into HassanAitKaci [9:33] AmandaVizedom: re: my 12:31 comment - should have said "offline" rather than "invisible." Right click on contact, choose "Call" then choose "Skype call." [9:33] anonymous morphed into Gary Gannon [9:35] anonymous2 morphed into Leora [9:35] anonymous morphed into Karl Hebenstreit [9:35] anonymous morphed into FrancescaQuattri [9:35] AliHashemi: joinconference is not online but invisibel? [9:35] AliHashemi: invisible* [9:36] SimonSpero: joinconference is hiding but working [9:36] AmandaVizedom: Ali, can you see my 12:31 and 12:33 comments above? [9:37] anonymous morphed into dennis pierson [9:38] anonymous morphed into JulienCorman [9:40] anonymous2 morphed into MichaelRiben [9:40] anonymous1 morphed into Brian Haugh [9:40] anonymous2 morphed into GenZou [9:40] AliHashemi: Just saw them, thanks. [9:41] anonymous1 morphed into PavithraKenjige [9:41] PavithraKenjige: Mute using *6 [9:45] PeterYim: == LeoObrst and Benjamin Grosof starts the session - please open the slides at: http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2013_10_24#nid3ZR3 [9:45] anonymous1 morphed into Beth [9:46] HensonGraves: I cannot get to the web page any more under any browser [9:49] PeterYim: == Opening Remarks: see - http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2013_10_24#nid3ZQT [9:49] PeterYim: @Henson - the wiki session page should be working properly (I just checked from my own browser) [9:53] anonymous1 morphed into FrancescaQuattri [9:53] PeterYim: those who do not have slides for their opening remarks are encouraged to capture their thoughts into this chat-room (or send me their slide(s) which I can add back to the archives) [9:56] PeterYim: @Leo, @Banjamin - (please remind the participants) we have 46 people on the phone-bridge, but only 38 in the chat ... please join us in the chat-room if you aren't already ... ref. details at top of session page [9:56] anonymous1 morphed into Paul Fodor [9:59] FrankOlken: Leora, Contact me about an upcoming workshop in DC Nov. 14-15 on information sharing for financial regulation. I think you would find it interesting. [10:01] anonymous morphed into Snezana Nikolic [10:02] List of members: Adrian Giurca, AlexShkotin, AliHashemi, AmandaVizedom, BenjaminGrosof, Beth, Beth1, BrandonWhitehead, Brian Haugh, ChristopherSpottiswoode, dennis pierson, ElieAbiLahoud, FrancescaQuattri, FrankOlken, Gary Gannon, GenZou, Harold Boley, HassanAitKaci, HensonGraves, JulienCorman, Karl Hebenstreit, KenBaclawski, LeoObrst, LeoObrst1, Leora, MBarnett, MichaelGruninger, MichaelRiben, MikeDean, NaicongLi, Paul Fodor, PeterYim, Richard Martin, SimonSpero, Snezana Nikolic, TaraAthan, Todd Pehle, ToddSchneider1, TonyWeida1, Weihong Song [10:03] Leora: @Vinay: The reason we need a rule language and not just ontologies, is that ontologies aren't sufficiently expressive. [10:03] Leora: We need rules with n-ary predicates, for examples. [10:03] Leora: We need to represent exceptions and default reasoning. [10:04] AmandaVizedom: +1 to remarks about artificiality of separating "rules" from "ontology." IMHO this creates artificial barriers and obstacles for projects using a representation that separates these. [10:04] Leora: We can't do all that we need in an ontology. [10:04] Leora: However, an ontology is an important component in our system. [10:04] SimonSpero: @Leora: that's crazy talk. Next you'll want default reasoning or something :-P [10:04] Leora: @Simon, indeed I do want default reasoning. And yes, crazy talk is what I do. [10:05] anonymous morphed into Frank Chum [10:05] AmandaVizedom: @Leora, I think the point is that there is no reason to think of the rules as something apart from the ontology. That separation is what gives us insufficiently expressive ontologies. But many ontologies/ representations include rules as an essential part of the ontology. [10:05] FrankOlken: Leora, I agree with the need for rules with n-ary predicates. [10:07] Leora: @Amanda, yes, some rules, but not all kinds of rules. Rules with n-ary predicates? default rules? [10:07] ToddSchneider1: Leora, Frank, The issue is the expressivity of representation language. Some are less expressive than others. [10:07] LeoObrst: The separation of ontology from rules is probably an artifact of semantic web technologies, i.e., DL-based OWL. Other ontology languages make no such separation. [10:07] Leora: @Todd, frankly, we need more than FOL. [10:08] AmandaVizedom: @Leora, that will vary with expressiveness of language. And different use cases will require different expressiveness. [10:10] Leora: @Amanda, certainly, if you expand your definition of ontology to include rules with n-ary predicates, default rules, modal operators (deontic logic), etc, then sure. But then we're just redefining the word, no? [10:10] ToddSchneider1: Leora, for your problem space, probably. The question(s) that should be asked is what problems are you trying to solve. This should then provide requirements for the needed expressivity and the language needed (and any supplements). [10:12] AmandaVizedom: @Todd, I think that's exactly right. What I took Vinay to be addressing is this (relatively recent) phenomenon in which people discover that limited-expressiveness languaguages, not supporting rules, are insufficient for their needs, and then create or adopt a *separate*, added-on language for rules. It seems much more sensible instead to move to a higher-expressiveness, rule-inclusive representation -- of which there are many, already developed and understood. [10:13] ToddSchneider1: Amanda, the largest constraint (from a production viewpoint) is the viability of the infrastructure to support the choice. [10:13] SimonSpero: Anyone here who knows anything about Typed Feature Structures and unification? :-P [10:14] AmandaVizedom: @Leora, for many of us who have been working in ontologies for many years, we are experiencing a redefinition in the other direction. This idea of ontologies as not including rules is the new one, and when it first started popping up in literature/ conferences, etc., it was quite puzzling. [10:15] Leora: @Amanda, I understand your point. But I think you can see where I'm coming from. It's a complex domain, and we need a very expressive representation, and I don't know of any ontology that gives what I need. If you can point one out, I'll be happy to look at it,. [10:15] Beth1: It might help if we are clear about the difference between the language, and the model that is the basis for assigning meanings to terms in the language. If you think the "ontology" doesn't include rules, that might be because you are thinking of the ontology as the model. But the language can surely include terms and functions and operators as needed for describing the domain. In that sense, there's no need to have separate languages for the "ontology" and the rules. I'm putting "ontology" in quotes because we are pretty fast and loose in the semantic tech world about what an ontology is. [10:16] anonymous morphed into Hamizah Hamka [10:16] AmandaVizedom: Todd, that's very true. The best-tooled, best-tested, higher-expressiveness systems are also mostly proprietary. Currently, most(?) or at least many projects require or prefer an open standard representation langugage, and we lag badly in the support for the higher-expressiveness options there. [10:18] FrancescaQuattri: @Harold. I am looking at your slides (no. 5). A question about the concept of "reaction rules". Did I get it right when you suggested that we could reuse the same ontological rules from an ontology to another in no random order, but according to the subsumed examples that you proposed? e.g. spatio ontology subsumes to temporal ontology, action ont. to event ont. etc.? Thank you for the explanation. [10:19] ToddSchneider1: Beth, an ontology is a model (of some sort). [10:20] ToddSchneider1: Amanda, Yes. Any large customer usually prefers or requires products that comply with [open] standards. [10:21] AmandaVizedom: @Leora, I definitely get it. And I started my ontology career working with Cyc / CycL, so was spoiled at the outset with respect to expressivity and the use of rules within ontology. I hope that in the process of this miniseries, we see more richness of options! [10:21] PeterYim: == LeoObrst presenting: "Survey: Logic, Logic Programming, Ontology, Rules" [10:23] List of members: Adrian Giurca, AlexShkotin, AliHashemi, AmandaVizedom, BenjaminGrosof, Beth1, BrandonWhitehead, Brian Haugh, ChristopherSpottiswoode, dennis pierson, ElieAbiLahoud, FrancescaQuattri, Frank Chum, FrankOlken, Gary Gannon, GenZou, Hamizah Hamka, Harold Boley, HassanAitKaci, HensonGraves, JoelBender, JulienCorman, Karl Hebenstreit, KenBaclawski, LeoObrst, LeoObrst1, Leora, MBarnett, MichaelGruninger, MichaelRiben, MikeDean, NaicongLi, Paul Fodor, PavithraKenjige, PeterYim, Richard Martin, SimonSpero, Snezana Nikolic, TaraAthan, ToddSchneider1, TonyWeida1, VinayChaudhri, Weihong Song [10:24] SimonSpero: Program = algorithm + data. Algorithm = logic + control . Program = logic + control + data . Ontology = program - control? [10:31] FrancescaQuattri: got the def. of reaction rules. thank you Leo [10:34] SimonSpero: Prolog is a high level WAM assembler language [10:38] BenjaminGrosof: What Leo is referring to with LP is often the LP family of KRs, incl. extensions for skolemization, defeasibility, Rulelog, etc. [10:41] Harold Boley: @Francesca, the "Reaction Rules" slide gives the big picture of Reaction RuleML. In the third row, spatio and temporal ontologies are often used together, as are action and event ontologies. But generally, (Reaction) RuleML provides a 'pluggable' architecture where you can modularly combine various ontologies. We are currently re-specifying Reaction RuleML 1.0 from XSD in Relax NG. This would allow MYNG-style customization of Reaction RuleML as we already do for Deliberation RuleML: http://wiki.ruleml.org/index.php/MYNG [10:42] anonymous morphed into Oscar [10:47] PeterYim: == BenjaminGrosof presenting: "Survey of Knowledge Representations for Rules and Ontologies" [10:51] PeterYim: ... on slide #3 now [10:55] AmandaVizedom: Regarding @BenjaminGrosof's slide 3 point 3: Agreed, but a notable point: anyone who takes even an Introduction to "Symbolic Logic" or "Formal Logic" course (usually taught in a Philosophy or Mathematics department, but often taken much more broadly) in college *does* learn the rules and key ingredients here. Test: if your Intro to Logic class had variables and Quantifiers, you learned this. :-) [10:56] SimonSpero: Imperial College still has mandatory Prolog (sequenced after the first Logic course) [10:56] Hamizah Hamka: Apologize team of experts.Need to leave too soon for its too late here. Hope to participate in future event. Have a fruitful discussion. [10:57] AmandaVizedom: +1 Slide 4: viewing as "rules" or "ontological statements" is largely a matter of *view* and *use*. In fact, expressive KRs include "syntactic sugar" to allow moving between the views for many kinds of information. [10:58] PeterYim: @Hamizah Hamka - thank for participating ... all the way from Malaysia (right?) [11:00] AmandaVizedom: Big point in support of the "rules" & "ontological knowledge" overlap: The fundamental ontological relationship of "subclass" or "subtype" is a (usually hard-coded) expression of a rule pattern: "A is a subclass of B" means: "If x is an instance of A then x is also an instance of B." It's that fundamental. [11:00] anonymous morphed into Alessandro Provetti [11:03] FrancescaQuattri: @Harold: Among the good things that are in the wiki (shared link) is the fact that it contains examples (ref. "Example Instance Files for RuleML 1.0"). It helps to frame the final intent into real world applications. [11:06] SimonSpero: I'm not sure that the OWL choice of OWA (with no Clark completion) keeps it in the mainstream of LP? [11:07] PeterYim: ... on slide #12 now [11:08] SimonSpero: ( http://www.doc.ic.ac.uk/~klc/neg.html ) [11:12] HassanAitKaci: Before Hilog: lambda-Prolog, and its later incarnation Teyjus ( http://teyjus.cs.umn.edu/ ) [11:17] PeterYim: ... on slide #18 now [11:17] PeterYim: == Q&A and open discussion ... please raise your hand, get called upon by the chair, test your voice, before making your remark [11:17] Harold Boley: You can practically explore higher-order Hornlog syntax as developed for Hilog (http://www.cs.sunysb.edu/~warren/xsbbook/node45.html) and Relfun (http://www.cs.unb.ca/~boley/FLP/cs6905FLP.pdf) using the online Relfun interpreter: http://www.dfki.uni-kl.de/~vega/relfun-cgi/cgi-bin/rfi.cgi [11:18] anonymous morphed into FrancescaQuattri [11:19] anonymous1 morphed into Alex@ITT exelis [11:19] HassanAitKaci: I'm on Skype and unmuted ... [11:19] HassanAitKaci: I don't understand [11:21] Harold Boley: Dale Miller's et al.'s lambda-Prolog actually covers a (second-order) fragment of higher-order logics, not just higher-order syntax: http://www.lix.polytechnique.fr/~dale/lProlog/ [11:24] PeterYim: @whoever just commented: "Can either presenter comment on the application of temporal descriptors as terms in logic programs?" ... please (re-)type your message into the box on the left of the "send" button [11:24] SimonSpero: WAM - A tutorial reconstruction - http://wambook.sourceforge.net/ [11:25] Alex@ITT exelis: Can either of the presenters comment on the application of tmeporal descriptors as terms? in logic programs? [11:25] LeoObrst: As mentioned by Hassan, efficiency is important, and such an area as knowledge compilation addresses how to "compile" very expressive knowledge into more efficient runtime representation. [11:28] SimonSpero: "With imperative programming, you have to tell the computer how to do what you want. With declarative programming, you have to trick the computer in to doing what you want." [11:28] JoelBender: Is there a description of RuleML and its component "ontologies" that describes it concepts in a way that is similar to FOAF or RDFS? [11:29] Alex@ITT exelis morphed into AlexMirzaoff@ITT exelis [11:29] AmandaVizedom: [I just briefly mis-heard "attribution" as "retribution" :-\!] [11:30] AlexMirzaoff@ITT exelis: Can either of the presenters comment on the application of tmeporal descriptors as terms? in logic programs? dynamic ontologies? [11:30] AmandaVizedom: @HassanAitKaci: Got it, thank you. The ')' got auto-incorporated into the link in your original post, that was all. [11:33] Harold Boley: Data models of XML (positional) and RDF (slotted) can be reconciled: http://www.dfki.uni-kl.de/~boley/xmlrdf.html [11:33] PeterYim: @Amanda - very funny! :) [11:34] AmandaVizedom: @Peter :-) [11:37] Harold Boley: RuleML uses XML in the object-oriented manner enabled by this reconciliation: Distinction of capitalized Nodes ('types') and lower-cased edge ('role') XML elmenets. [11:37] PeterYim: @JoelBender - can your type out your question (to which HaroldBoley) is answering, so we would have captured the context into the transcript [11:38] AmandaVizedom: Last year's "Best Paper Award" recipient at the the STIDS conference presented a particular ontological representation of time that supported a variety of reasoning approaches that have been often addressed in ways that have some built-in problems... [11:39] PeterYim: great kick-off, Leo & Benjamin! [11:39] HassanAitKaci: http://cs.brown.edu/people/pvh/CPL/Papers/v1/hak.pdf [11:39] PeterYim: join us again next week 2013_10_31 - Thursday: RulesReasoningLP mini-series session-02: Concepts and Foundations of Rules and Ontologies: Logic Programs, Classical Logic, and Semantic Web - I - Co-chairs: LeoObrst & HaroldBoley - watch out for announcement on the mailing list and the developing session page at: http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2013_10_31 [11:39] FrancescaQuattri: Thank you All! [11:39] AlexShkotin: Good bye [11:39] LeoObrst: Thanks, all! [11:40] SimonSpero: Good session. [11:40] AlexMirzaoff@ITT exelis: thanks Amanda, I will check that reference [11:41] AmandaVizedom: @Alex: see Shrag "Best-practice..." at http://stids.c4i.gmu.edu/STIDS2012/agenda2012.php [11:42] Harold Boley: The RuleML Normalizer RON can automatically generate object-oriented stripes/slots: http://wiki.ruleml.org/index.php/Specification_of_Deliberation_RuleML_1.0#XSLT-Based_Normalizer [11:42] AlexMirzaoff@ITT exelis: Amanda - got it thanks [11:45] AmandaVizedom: yw! [11:46] AmandaVizedom: Thanks all for this kick-off. Looking forward to the mini-series, and glad these matters are getting some deserved and needed attention.[11:40] PeterYim: -- session ended: 11:30am PDT -- [11:40] List of attendees: Adrian Giurca, Alessandro Provetti, Alex@ITT exelis, AlexMirzaoff@ITT exelis, AlexShkotin, AliHashemi, AmandaVizedom, BenjaminGrosof, Beth, Beth1, BrandonWhitehead, Brian Haugh, ChristopherSpottiswoode, ElieAbiLahoud, FrancescaQuattri, FrancescaQuattri1, Frank Chum, FrankOlken, Gary Gannon, Gary Gannon1, GenZou, Hamizah Hamka, Harold Boley, HassanAitKaci, HensonGraves, HensonGraves1, HensonGraves2, JoelBender, JulienCorman, Karl Hebenstreit, KenBaclawski, LeoObrst, LeoObrst1, Leora, MBarnett, MBarnett1, MichaelGruninger, MichaelRiben, MikeDean, NaicongLi, Oscar, Paul Fodor, PavithraKenjige, PeterYim, Richard Martin, SimonSpero, Snezana Nikolic, TaraAthan, Todd Pehle, ToddSchneider, ToddSchneider1, TonyWeida, TonyWeida1, VinayChaudhri, Weihong Song, Weihong Song1, anonymous, anonymous1, anonymous2, dennis pierson, dennis pierson1, vnc2 ------