From OntologPSMW

Revision as of 23:03, 15 October 2019 by KennethBaclawski (Talk | contribs)

Jump to: navigation, search
[ ]
    (1)
Session Introductory
Duration 1 hour60 minute
3,600 second
0.0417 day
Date/Time September 25 2019 16:00 GMT
9:00am PDT/12:00pm EDT
5:00pm BST/6:00pm CEST
Convener Ken Baclawski

Contents

Agenda     (2A)

Conference Call Information     (2B)

Participants     (2C)

Proceedings     (2D)

[12:23] RaviSharma: John described web, RDBMS, and precise ontology     (2D1)

[12:25] RaviSharma: janet - said about pragmatics     (2D2)

[12:26] RaviSharma: John said it is about context     (2D3)

[12:28] RaviSharma: Tim Berners Lee - WWW 1994 suggested Semantic Web     (2D4)

[12:29] RaviSharma: Jim Handler - DARPA came for 2yr and proposed the DARPA Agent Markup Language (DAML) project     (2D5)

[12:31] RaviSharma: There was split between CL and DL people     (2D6)

[12:32] RaviSharma: Polynomials > 2 does not scale to web     (2D7)

KenBaclawski: Actually, DL is worse than polynomial. It is NEXP, Nondeterministic Exponential.     (2D8)

[12:32] RaviSharma: RDF and OWL were only preliminary but people took it as rigid.     (2D9)

[12:36] RaviSharma: Guha student of John McCarthy disagreed with Lenat, CyC but OWL dominated     (2D10)

[12:37] RaviSharma: FOL powerful but can not do many things     (2D11)

[12:37] RaviSharma: SQL is FOL, OWL is inefficient     (2D12)

[12:38] RaviSharma: Prolog is faster even in executing OWL     (2D13)

[12:40] RaviSharma: Ken worked on DAML, but his suggestions which would have made the language equivalent to FOL where not accepted.     (2D14)

KenBaclawski: Pat Hayes was also part of the DAML project, but was not involved one way or the other in this decision, at least not at the time. I also suggested inclusion of a framework for limited contexts (i.e., situation semantics) but this was also not accepted.     (2D15)

[12:43] janet singer: From the Ontology Summit 2020 Intro: The theme of the Summit is to examine KGs from a number of points of view ranging from low-level representation and storage techniques to high-level semantics, and from the vendors to the end users.     (2D16)

[12:44] RaviSharma: Formal Concept Analysis (FCA) - simple attributes - monatic relation, create lattice, wordnet, thesaurus, multi inherited, by push of button, little or no knowledge for data inputs, Tim and Jim wanted that kind of data input and application.     (2D17)

[12:46] janet singer: A goal for the communiqué could be to produce graphics relating those multiple dimensions listed to get the big picture     (2D18)

[12:46] RaviSharma: DoD did not fund more. A lot more has been developed in KB since then, but the technology has stayed the same.     (2D19)

[12:48] RaviSharma: Janet suggested diff depictions and graphics, microtheories, small representations, benefits of visuals, till DOL site as janet suggested, essentially look at where KGs sit     (2D20)

[12:49] RaviSharma: For the communiqué an overview diagram showing many of these elements.     (2D21)

[12:49] janet singer: HETS site, John's diagrams under IKL as reference points     (2D22)

[12:50] RaviSharma: John- Oracle tried to block SQL open standards and weakened interoperability     (2D24)

[12:52] RaviSharma: Jim's diagram, Tim Berners Lee both diagrams, many are on John's URL we can borrow them for overall communiqué     (2D25)

[12:53] RaviSharma: John's web page on Semantic Networks for Wiley updated but KG to be added     (2D26)

[12:53] RaviSharma: add KGs     (2D28)

[12:54] RaviSharma: CGIF and KG Q from? KG is a subset     (2D29)

[12:55] RaviSharma: John said existential and other kinds of graphs, so CL is base for representing all this     (2D30)

[12:55] RaviSharma: CGIF is linearization EGIF nd RDF can also be mapped to that     (2D31)

[12:58] RaviSharma: Arun Majumdar was not able to login he could in January or February may speak. John will give an introduction and Arun will describe software     (2D32)

[12:59] RaviSharma: Unnatural languages? Formal languages like software languages John says they can handle incompatibilities.     (2D33)

[13:01] RaviSharma: The least common ontology and parts in or not in common give you framework - go up lattice for lowest commonality     (2D34)

[13:02] RaviSharma: standards - dictionaries - cases - janet     (2D35)

[13:02] Douglas R. Miles: https://grakn.ai/slack <- link to developer chat on Grakn     (2D36)

[13:03] RaviSharma: Pronunciations change relatively slowly. Meanings change and are added more quickly.     (2D37)

[13:05] RaviSharma: Cognitive memory attempts to map computer language to NL reverse engineering - all software with documents data and email - analysis of all knowledge in their software repository     (2D38)

[13:08] Douglas R. Miles: Ken, that link is how I usually contact Ganeshwara Herawan Hananda     (2D39)

[13:08] RaviSharma: John is saying all this - 40 person-years was the Accenture estimate, but Arun and Andre completed it in a 15 person-weeks project. They did semi-automatic Vivomind and complete analysis of JCL COBOL, engineering comments, parsed grammar with Prolog; using these derived an ontology from COBOL; mapped the English ontology; and then achieved a richer result in 15 weeks.     (2D40)

[13:08] RaviSharma: Arun and Andre achieved it.     (2D41)

[13:09] RaviSharma: Formal computer language to English NL.     (2D42)

[13:09] RaviSharma: John's slides showed this.     (2D43)

[13:12] RaviSharma: microfocused on COBOL for Unix, Ravi used it     (2D44)

[13:14] RaviSharma: Jeopardy Watson are making dent -John     (2D45)

[13:16] RaviSharma: John -Allegrograph, Franz, version of Prolog all is niche market, there is immense market for Apps     (2D46)

[13:16] RaviSharma: Continue with discussion next week     (2D47)

KenBaclawski: Note that there will not be a session two weeks from today (9 October).     (2D48)

[13:18] RaviSharma: Doug Miles spoke also     (2D49)

Resources     (2E)

Previous Meetings     (2F)


Next Meetings     (2G)