Actions

Ontolog Forum

Revision as of 06:08, 9 January 2016 by imported>KennethBaclawski (Fix PurpleMediaWiki references)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

Ontolog Forum Panel Discussion Preparation Page

Thursday 20 April 2006

  • Topic: "Ontologizing the Ontolog Body of Knowledge - Discussion Session 1 - Framing the Issues, Requirements and Approach"

Genesis and Background of this discussion

Proposed discussion agenda

  • Session Format: this is be a virtual session conducted over an augmented conference call
    • 1. we'll go around with a self-introduction of participants (10~15 minutes)
    • 2. Introduction of Panelists (Moderator)
    • 3. Opening by the Moderator (5 min.)
    • 4. 5-minute summary presentation by each panelist on their perspective
    • 5. follow each panelist's presentation with a 3 minute Q&A and discussion on that particular perspective
    • 6. open general discussion by all participants
    • 7. Summary / Conclusion / Follow-up (Moderator)

Pertinent Issues

(please draft and list)

  • Framing the Issues
    • Some 'framing questions': (--DeniseBedford / 2006.04.17-06:57 EDT)
      • What content from Ontolog should be ontologized? (content defined broadly - people, groups, email messages, conference call agendas, conference call transcripts/recordings, presentations, discussion threads, etc.)
      • What content from outside Ontolog should be ontologized? (content referenced but not created by the group...?, standards, conferences, research & development, other communities)
      • At what level do we want to apply the ontology to the content?
      • Can we define a set of roles or user types of the Ontolog content - from within Ontolog and external to Ontolog?
      • Can we define a handful of scenarios to use to test the ontology development as we move forward?
      • How will the ontology be used on a practical level?
      • What applications will consume it?
        • Will we use it in internal architecture/applications only?
        • Will we surface portions of it for consumption by external applications?
      • What kind of an underlying functional architecture do we need?
        • What kinds of applications and technologies will support this functional architecture?
  • Establishing Requirements
  • Approach the Task

Panelist Focus

Speaker - Title of Brief - Abstract

  • Denise Bedford -
    • (abstract) -
  • 0. Peter P. Yim - "The Ontolog Collaborative Work Environment - Our Community Knowledge Repository"
    • (abstract) - Practically all of Ontolog's electronic transactions have been, for the last 4 years, captured in the hosted collaborative work environment ("CWE") at <http://ontolog.cim3.net>. This panel member will briefly introduce to the participants how the CWE is structured and organized. He will further expose the key tenets of the CWE design, and how that has been inspired and influenced by the Doug Engelbarts notions of "Bootstrapping" and prompts that we should look at more than just "use cases" in developing our approach and requirements.
  • 1. Lisa Colvin - (Use and Context: understanding user's role and view on the content and the content itself ... taking it from experience.)
    • (abstract) - I'll be referencing Denise's presentation 7-13. Also will take into consideration Pat's suggestion of different phases of membership (from passive (listening/learning) to active)
  • 2. Bob Smith - A roadmap for Ontolog's Thesaurus: Sketch 1
    • (abstract) - Denise Bedford presented a framework for us to Ontologize Ontolog's wiki-web content reguires many different but orchestrated tasks, including a Thesaurus.
She has provided an initial inventory of terms (Concepts and Topics) on which to build the architecture needed to achieve this stated goal ( OO).
Building a Thesaurus which fits with other evolving Ontolog Forum elements suggests a basic blueprinting of process based upon a roadmap that shows the various feasible paths between nodes/sub-goals. Feasible paths avoid known and reported hazardous areas of endless debate, frustrating loops, and sink-holes frequently found in academic and/or philosophical sqabbles over symbols without real world reference. (See "the Meaning of Meaning" debates and the Semantic Triangle linking concepts, symbols, reality checks. )
A driver of this framework is the need for Applied Ontology, as defined by the journal ( http://www.iospress.nl/html/15705838.php ) and frequently by this Ontolog Forum. (--BobSmith / 2006.04.17 8:56am PDT)
  • 3. Patrick Heinig - (speak to the range of tools we might need to move the project forward ... value proposition)
    • (abstract) -
  • 4. Nicolas Rouquette - (speak to functional architecture)
    • (abstract) - Nicolas: want to know (a) what we want to tag, (b) how to classify (e.g. some architecture, UIMA, etc.), (c) what tools are available to help us do it, and (d) how to extract more knowledge from the tagged content   
  • 5. EMichaelMaximilien - Multifaceted approach to ontologizing the ONTOLOG content rooted in

pragmatism, tools, services, and standards, and social collaboration (--EMichaelMaximilien / 2006.04.18-10:42 PDT; updated 2006.04.18-18:27 PDT)

    • (abstract) - In trying to "ontologize" the ONTOLOG content we must not forget two of the key criticisms and issues with ontologies: (1) lack of pragmatism in the goals and applications of ontologies; and (2) heterogeneity of usage and use cases. On the one hand the ontologies that we would associate with the community's content should stem from the pragmatic applications that are intended (the use cases) and on the other hand, we should assume that these use cases are many and various. Using this thesis and assertions, I propose that we take a multifaceted approach to our problem.
First, simple tagging, a la http://del.icio.us, can be used to allow individual users to tag the content as they view it in an idiosyncratic manner. Assuming enough such tagging and by different users, we would create folksonomies [0] [1] for the content. We can complement the tagging with other social features such as comments and community feedback (i.e., ratings, digging, mojo) as done in Web 2.0 services, e.g., http://digg.com and http://slashdot.org, which would help involve users and give a sense for the content's value in the eyes of its users.
Second, some of the main content are in the form of audio files, the approach there is to use services such as http://podzinger.com and others, to convert the audio into text transcripts and tags.
Third, as leading experts in the field of ontology engineering, we should be able to fight our way into creating at an initial ontology for the site's content. This ontology would give us an outline for the site.
And finally, using mining tools, standards, and techniques, such as UIMA [2] we can get an initial automated annotation of the content.
With these four different sources of semantics, the directions and questions then become what are the use cases we are trying to solve and how do we combine these semantic sources (with appropriate algorithms, UI) to solve the problems posed by the use case?
For instance, cataloguing could be a combination of using the folksonomies [3] and automated annotations to associate the content with appropriate concepts with the initial outline ontology. Social features for rating and allowing feedback (similar to http://digg.com and http://slashdot.com) would help in the contents' presentation by helping infer its utility.

[0] http://en.wikipedia.org/wiki/Folksonomy

[1] http://tomgruber.org/writing/ontology-of-folksonomy.htm

[2] http://www.research.ibm.com/UIMA/

[3] Mika, P. "Ontologies are us: A unified model of social networks and semantics", ICSW 2005, Galway, Ireland

Pertinent Questions for Discussion

(please draft and list)

  • For the Panelists
    • one of us should review the Methodology the community discussed established for the UBL-Ontology project back in 2003, and do a compare and contrast between "that" and "this", and re-use the lessons learned
    • enumerate the goals of this entire exercise (=ppy/2006.04.15-09:05PDT)
    • enumerate the skill sets that will need to be engaged if we were to do a good job at it (=ppy/2006.04.15-09:05PDT)
  • For the participants
    • what aspects of Ontolog (it's community, body of knowledge, people, interaction, choice of issues it tackles, ... ) do you find to be of value? (=ppy/2006.04.15-10:48 PDT)
    • name the top few things that you would want to be able to do with Ontolog (=ppy/2006.04.15-09:05PDT)
    • Are there any communities that you feel we should be including in our work? Did your needs feel addressed when you first joined the community?
    • name one or two things you had wished you were able to do (find at) at Ontolog, but was not
    • since what we can deliver s totally dependant on community participation, please provide some suggestions as to how we can get the right people (with the right skill set) engaged (=ppy/2006.04.15-09:05PDT)

File Repository

Logistics

  • Kurt Conrad has agreed to take note on the wiki during the 2006.04.20 session