Hello, I am a student in computer sciences in France and I am studying argumentation in logic to represent reasoning and decision making.
To do experimentation, I am using an implementation in prolog of an argumentation framework that represents rules by classical horn clauses, and that let specify preferences on contradictory rules, and preferences on contradictory preferences (high order pref). A classical example of this is the tweety example: r1 : fly(X) <- bird(X) r2 : neg(fly(X)) <- penguin(X) f1 : bird(tweety) f2 : penguin(tweety) Like this we can infer fly and neg(fly) for tweety. The framework let us tell: pr1 : h_p(r2, r1) (h_p as higher preferency) And so we can say that r2 is a particular case of r1 and things like that. Another interesting things that can be done with gorgias is to extract the "explanation" of why a goal is proved, so here we would have had : [penguin(tweety), pr1] as an explanation. The implementation can be found here http://www2.cs.ucy.ac.cy/~nkd/gorgias/ where there is more information on the possibilities of it (there is a tutorial with multiple examples). I am highly interested in Oz and Mozart and I would like to "port" gorgias or something similar that let programmer specify an argumentation theory (with rules and preferences, negation, abducibility and high order preferences) to Oz. Since I have not very much experience in languages like Oz, and since Oz is a lot different that prolog (specially for search), I am wondering what is the best way to do it. I need to : 1) represent rules like the tweety's ones. 2) be able to have access to them like in prolog to compare their head, prove their body. What could be interesting is to have the rules as procedures, like if I was converting prolog code to Oz, but to be able to do a "special" search on them that would be able to see if they are conflictual, remember the path taken, treat specially a special king of predicate (for abducibility), have a way to specify preferences, and a have a body for these preferences that could be searched too ... I read a lot and I have seen that I could make my own search procedure, but I am not sure how I can store the rules and preferences, extract the explanation, specify that special predicates are abducibles, specify negation, etc ... I hope it is a subject that will interests people, since I began to play with argumentation, I could not stop :) Thanks, Victor
signature.asc
Description: Digital signature
_________________________________________________________________________________ mozart-users mailing list [email protected] http://www.mozart-oz.org/mailman/listinfo/mozart-users
