Adhominem Comments http://adhominem.blogsome.com Argumentation: logic, theory & implementation .... and the tricky line between them Wed, 11 Nov 2009 19:03:33 +0000 http://wordpress.org/?v=1.5.1-alpha by: Terry http://adhominem.blogsome.com/2007/10/03/integrating-ontologies-argumentation-for-decision-making-in-breast-cancer/#comment-10 Wed, 07 Nov 2007 04:44:58 +0000 http://adhominem.blogsome.com/2007/10/03/integrating-ontologies-argumentation-for-decision-making-in-breast-cancer/#comment-10 i need your investigate i need your investigate

]]> by: Annerose http://adhominem.blogsome.com/2006/01/25/ontologies-folksonomies-etc/#comment-9 Sun, 03 Jun 2007 13:01:17 +0100 http://adhominem.blogsome.com/2006/01/25/ontologies-folksonomies-etc/#comment-9 These comments have been invaluable to me as is this whole site. I thank you for your comment. These comments have been invaluable to me as is this whole site. I thank you for your comment.

]]> by: Rosie http://adhominem.blogsome.com/2006/02/03/ontology-vs-inference-in-logic/#comment-8 Sat, 28 Apr 2007 15:28:52 +0100 http://adhominem.blogsome.com/2006/02/03/ontology-vs-inference-in-logic/#comment-8 These comments have been invaluable to me as is this whole site. I thank you for your comment. These comments have been invaluable to me as is this whole site. I thank you for your comment.

]]> by: Dan Brickley http://adhominem.blogsome.com/2006/01/25/ontologies-folksonomies-etc/#comment-7 Mon, 19 Feb 2007 23:58:01 +0000 http://adhominem.blogsome.com/2006/01/25/ontologies-folksonomies-etc/#comment-7 An additional point, is that W3C's approach - with RDF, RDFS and OWL - is fundamentally pluralistic by design. We created a system for multiple independently developed ontologies to be mixed-and-matched according to need and whim. The idea that an ontological approaches forces a single rigid view, ... is inaccurate w.r.t. the Web. Also, there are "scruffy" ontologies like SKOS that allow simple hierarchical concept schemes to be reflected into RDF. An additional point, is that W3C’s approach - with RDF, RDFS and OWL - is fundamentally pluralistic by design. We created a system for multiple independently developed ontologies to be mixed-and-matched according to need and whim. The idea that an ontological approaches forces a single rigid view, … is inaccurate w.r.t. the Web. Also, there are “scruffy” ontologies like SKOS that allow simple hierarchical concept schemes to be reflected into RDF.

]]> by: Karen http://adhominem.blogsome.com/2006/01/27/shirky-and-ontology-vs-folksonomies/#comment-4 Wed, 01 Feb 2006 21:43:45 +0000 http://adhominem.blogsome.com/2006/01/27/shirky-and-ontology-vs-folksonomies/#comment-4 Random thoughts on the n% belonging idea: it seems to be that there are 2 different ways in which something can fail to always fit into a certain category, and to have a percentage associated with it. On the one hand, you can have something that occasionally matches category A but then occasionally matches category B, with a complete transition between the two categories. In that case, you could estimate the amount of time in which the item was category A versus category B and say that something is 60% A and 40% B. More problematic, though, would be things that share a percentage of characteristics with category A and a percentage with B. Then, if you say that something is 60% A and 40% B, you means something different: you mean that it has combination of characteristics from both categories. Both approaches might be useful, but troublesome to work with. Also, I think even Shirky would agree with you that "throwing ontology out the window" is not a necessary step. In the article on classification, he points out that certain domains (with expert users, formal categories, established boundaries, etc) still lend themselves well to the ontological approach. From what I can gather from your blog here, it sounds like the project your working on tends more towards that kind of domain than to the sprawling chaos of web communities. Apologies if this comment ends up doubleposted. I tried once and it didn't seem to take... Random thoughts on the n% belonging idea: it seems to be that there are 2 different ways in which something can fail to always fit into a certain category, and to have a percentage associated with it. On the one hand, you can have something that occasionally matches category A but then occasionally matches category B, with a complete transition between the two categories. In that case, you could estimate the amount of time in which the item was category A versus category B and say that something is 60% A and 40% B. More problematic, though, would be things that share a percentage of characteristics with category A and a percentage with B. Then, if you say that something is 60% A and 40% B, you means something different: you mean that it has combination of characteristics from both categories. Both approaches might be useful, but troublesome to work with.
Also, I think even Shirky would agree with you that “throwing ontology out the window” is not a necessary step. In the article on classification, he points out that certain domains (with expert users, formal categories, established boundaries, etc) still lend themselves well to the ontological approach. From what I can gather from your blog here, it sounds like the project your working on tends more towards that kind of domain than to the sprawling chaos of web communities.
Apologies if this comment ends up doubleposted. I tried once and it didn’t seem to take…

]]> by: Karen http://adhominem.blogsome.com/2006/01/27/shirky-and-ontology-vs-folksonomies/#comment-3 Wed, 01 Feb 2006 21:40:51 +0000 http://adhominem.blogsome.com/2006/01/27/shirky-and-ontology-vs-folksonomies/#comment-3 Random thoughts on the n% belonging idea: it seems to be that there are 2 different ways in which something can fail to always fit into a certain category, and to have a percentage associated with it. On the one hand, you can have something that occasionally matches category A but then occasionally matches category B, with a complete transition between the two categories. In that case, you could estimate the amount of time in which the item was category A versus category B and say that something is 60% A and 40% B. More problematic, though, would be things that share a percentage of characteristics with category A and a percentage with B. Then, if you say that something is 60% A and 40% B, you means something different: you mean that it has combination of characteristics from both categories. Both approaches might be useful, but troublesome to work with. Also, I think even Shirky would agree with you that "throwing ontology out the window" is not a necessary step. In the article on classification, he points out that certain domains (with expert users, formal categories, established boundaries, etc) still lend themselves well to the ontological approach. From what I can gather from your blog here, it sounds like the project you are working on tends more towards that kind of domain than to the sprawling chaos of web communities. Random thoughts on the n% belonging idea: it seems to be that there are 2 different ways in which something can fail to always fit into a certain category, and to have a percentage associated with it. On the one hand, you can have something that occasionally matches category A but then occasionally matches category B, with a complete transition between the two categories. In that case, you could estimate the amount of time in which the item was category A versus category B and say that something is 60% A and 40% B. More problematic, though, would be things that share a percentage of characteristics with category A and a percentage with B. Then, if you say that something is 60% A and 40% B, you means something different: you mean that it has combination of characteristics from both categories. Both approaches might be useful, but troublesome to work with.
Also, I think even Shirky would agree with you that “throwing ontology out the window” is not a necessary step. In the article on classification, he points out that certain domains (with expert users, formal categories, established boundaries, etc) still lend themselves well to the ontological approach. From what I can gather from your blog here, it sounds like the project you are working on tends more towards that kind of domain than to the sprawling chaos of web communities.

]]> by: Bijan Parsia http://adhominem.blogsome.com/2006/01/25/workdiary25106/#comment-2 Mon, 30 Jan 2006 05:09:17 +0000 http://adhominem.blogsome.com/2006/01/25/workdiary25106/#comment-2 I'm confused by your swoop comment. Swoop isn't directly dealing with Prolog. We do have an experimental version of Swoop with rule editing support which can be connected to an experimental version of Pellet that has AL Log support but needs SWI Prolog for the datalog reasoning. We hope soon to have support for DL Safe rules (see KAON2 for current support) and full SWRL. As for Argumentation, we do a lot of different sorts of analysis of OWL KBs for partitioning and debugging. The debugging/explanation stuff might be useful (as it identifies premises sufficient for an entailment). (I don't see Jena as supporting ontology and rules. Anyway, if I understood your requirements better, we might be able to help out.) I’m confused by your swoop comment. Swoop isn’t directly dealing with Prolog. We do have an experimental version of Swoop with rule editing support which can be connected to an experimental version of Pellet that has AL Log support but needs SWI Prolog for the datalog reasoning. We hope soon to have support for DL Safe rules (see KAON2 for current support) and full SWRL.

As for Argumentation, we do a lot of different sorts of analysis of OWL KBs for partitioning and debugging. The debugging/explanation stuff might be useful (as it identifies premises sufficient for an entailment).

(I don’t see Jena as supporting ontology and rules. Anyway, if I understood your requirements better, we might be able to help out.)

]]>