Few existing argumentation frameworks are designed to deal with probabilistic knowledge, and none are designed to represent possibilistic knowledge, making them unsuitable for many real world domains. In this paper we present a subjective logic based framework for argumentation which overcomes this limitation. Reasoning about the state of a literal in this framework can be done in polynomial time. A dialogue game making use of the framework and a utility based heuristic for playing the dialogue game are also presented. We then show how these components can be applied to contract monitoring. The dialogues that emerge bear some similarity to the dialogues that occur when humans argue about contracts, and our approach is highly suited to complex, partially observable domains with fallible sensors where determining environment state cannot be done for free.
|Title of host publication||Proceedings of the Twentieth International Joint Conference on Artificial Intelligence|
|Subtitle of host publication||Hyderabad, India, 6 - 12 January, 2007|
|Editors||Manuela M. Veloso|
|ISBN (Print)||9781577352983, 157735298X|
|Publication status||Published - 2007|
- Artificial intelligence -- Congresses