2 citations found. Retrieving documents...
Jaeger, M., Representation Independence of Nonmonotonic Inference Relations, in: Proceedings of KR-96, 1996.

 Home/Search   Document Details and Download   Summary   Related Articles   Check  

This paper is cited in the following contexts:
In Defence of the Maximum Entropy Inference Process - Paris, Vencovská (1997)   (Correct)

....cross entropy it is possible under certain circumstances to provide inference procedures which are closed under reasonable families of representation shifts. We refer the reader to [3] for details. A second, recent, remedy for the paradox of representation dependence is given by Jaeger in [4]. As in [3] Jaeger s inference processes, or measure selection functions as he calls them, are only required to pick a set of probability functions (rather than a unique probability function) satisfying the constraints. Jaeger gives a formalisation of what it means for a measure selection ....

....be I(S ) where S is the simplest set of constraints from which S could arise by a change of representation. Actually, in the case considered earlier of S = S 2 CL; g ME(S) is the set of all probability functions on SL, so the problem never even arises. We refer the interested reader to [4]. For a rather more radical solution, this time involving imprecise probabilities, see [5] We now suggest an alternative explanation as to why the source of may feel it to be paradoxical. This is that the speaker is not saying that s he was aware all along that A 1 was actually a disjunction ....

Jaeger, M., Representation Independence of Nonmonotonic Inference Relations, in: Proceedings of KR-96, 1996.


Measure Selection: Notions of Rationality and Representation.. - Jaeger (1998)   Self-citation (Jaeger)   (Correct)

....derived from a general minimum information principle, center of mass inference might be justified by a no less viable likelihood of evidence principle. In the second part of the paper we discuss the problem of representation dependence, both of center of mass and of maximum entropy inference. In (Jaeger 1996) a method was proposed that transforms selection rules I into variants I that are representation independent. In the current paper, results are presented that show what useful properties of I (particularly as expressed by rationality principles) we have to trade in for representation ....

....and Koller (1995) give a semantic definition of representation independence, based on embeddings f : A B of state spaces. This definition subsumes the atomicity principle, but can also be applied to selection problems not framed within the context of a propositional probabilistic logic. In (Jaeger 1996) a generalization of atomicity along a different line is provided by developing a concept of representation independence for arbitrary nonmonotonic logics. That concept, like atomicitiy, is syntactic, based on interpretations between formal languages. In the remainder of this section, a ....

[Article contains additional citation context not shown here]

Jaeger, M. (1996), Representation independence of nonmonotonic inference relations, in `Proceedings of KR'96', pp. 461-472.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC