Joshua Shinavier Laying the foundations for a World Wide Argument Web

From Semantic Portal Wiki

Presentation given at CSCI 6966 Advanced Semantic Web (Fall 2008)#Lesson 10

Presentation slides: Image:ASW-Joshua_Shinavier-presentation3.ppt

Note: the PDF of the article should be available through the ScienceDirect page below when accessing it through RPI's network. Otherwise, the article may only be available for purchase.

• Speaker: Joshua Shinavier
• Title: Laying the foundations for a World Wide Argument Web
• Authors: I. Rahwan, F. Zablith, and C. Reed
• Conference: Artificial Intelligence 171 (2007)
• URL: http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6TYF-4NPG0FD-1&_user=10&_rdoc=1&_fmt=&_orig=search&_sort=d&view=c&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=bc463728d918e523b24d8d92adabf78e
• Date of Presentation: 2008/11/06

Questions

ID	Question	Name	Answer
Joshua Shinavier Laying the foundations for a World Wide Argument Web GTW 1	Section 5.2 points out that one limitation of using RDFS for modeling is the lack of support for disjointedness (in defining "all classes of edges and nodes). Is there any reason why the authors might have chosen only to use RDFS when OWL would allow the use of disjoint declarations as well as using proper restrictions on the domain and range of properties (for example, disallowing edges between I-nodes). I find this choice particularly confusing since the sample RDF in section 5.2 uses a predicate a:minCardinality which presumably refers to the OWL term (but isn't discussed afaik). Moreover, the use of RQL (mentioned in section 6.2) as an RDF query language in a paper published in 2007 is rather surprising, and would seem to suggest that this work is quite a bit more dated than the publication date would indicate. Do you think the model described would benefit much from the use of more recent (read: appropriate) technologies such as OWL (and perhaps SPARQL), or are they mostly irrelevant implementation details?	Gregory Todd Williams
Joshua Shinavier Laying the foundations for a World Wide Argument Web Jesse Weaver	Out of curiosity, why didn't the authors use OWL for their ontology? They model their own cardinality semantics which could be replaced (granted, with different usage) by standard OWL constructs. If you look at their ontology (http://www.argdf.org/source/ArgDF_Ontology.rdfs), it shows that they've also defined their own InverseProperty semantics. (As a side note, I was surprised to find so many syntactical errors in the RDF/XML snippets for a journal article.)	Jesse Weaver
Rahwan2007laying question 1 by lebo	The authors chose to model their argument ontology in RDFS. Although this makes sense because of RDF's suitability for distributed augmentation of existing resources, how are the semantics of RDFS beneficial in this application? The authors restrict themselves to RDFS while wishing for disjointness. Why, in 2007, would an RDFS ontology developer choose not to use OWL? "Translating the ontology to more expressive Semantic Web ontology languages such as OWL can also enable ontological reasoning over argument structures, for example, to automatically classify arguments, or to identify semantic similarities among arguments." Could you suggest any examples for how (and why) arguments could be classified or similarities could be determined? The interfaces shown indicate a large jump from today's blogging to tomorrow WWAF. What kind of users would use the system the paper proposes? How much "argument theory" would they be expected to know? If a novice to "argument theory" used the system, how likely would their products align with the reasoning the system can provide? The system proposed involves adding additional explicit structure to the "blogging" of today. How far would a user need to decompose the free text, and how often would that granularity need to change as the argument evolves? At what point does the structure inhibit the thought process of the user?	Tim Lebo

Attendees

Tim Lebo