S.-W. Tan and J. Pearl. Qualitative decision theory. In Proceedings of the Thirteenth National Conference on Artificial Intelligence (AAAI'94), pages 928--932, 1994.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....specification. For example, we might lose the ability to quickly recognize when dominance arguments render detailed utility calculations redundant. These difficulties have been one of the motivations for interest in qualitative theories of probability and utility (i.e. preference) e.g. [von72, Bou94, Pea93, TP94a, TP94b, DW91, DW94, DSW91]. In the following sections we will discuss two recent papers of ours [BG95, BG96] which have the common theme of trying to apply various independence concepts (from the field of multi attribute utility theory) to decision making. We have two motivations for pursuing this approach. Our primary ....

.... very large , and all we care about are order of magnitude distinctions) one can hope to simplify the reasoning process. There are many interesting variants of this basic idea, including the use of qualitative probabilities alone (such as rankings [Pea93] or qualitatively ranked utilities alone [TP94a]. The above notions of utility independence can be applied mutatis mutandis to qualitative utility functions so defined, and as in the purely numeric case they continue to provide a useful means of specifying additional structure. For reasoning about probabilities, graphical models which capture ....

[Article contains additional citation context not shown here]

S. Tan and J. Pearl. Qualitative decision theory. In Proc. 12th National Conference on Artificial Intelligence (AAAI '94), pages 928--932, 1994.

....with beliefs and preferences that do not necessarily conform to the axioms of classical decision theory. Examples of work along these lines include Boutilier s work on the representation of and reasoning with qualitative statements of preference and normality [5] similar work by Tan and Pearl [43] on specifying and querying statements of conditional preference, the work of Dubois and Prade on possibilistic analogous of classical decision theoretic tools [16, 17] and Lehmann s work on more qualitative versions of Savage s theorem [34] Our work, too, ts within this approach since it ....

S. Tan and Judea Pearl. Qualitative decision theory. In Proceedings of the Twelfth National Conference on Arti cial Intelligence, pages 928-932, 1994.

....with beliefs and preferences that do not necessarily conform to the axioms of classical decision theory. Examples of work along these lines include Boutilier s work on the representation of and reasoning with qualitative statements of preference and normality [5] similar work by Tan and Pearl [43] on specifying and querying statements of conditional preference, the work of Dubois and Prade on possibilistic analogous of classical decision theoretic tools [16, 17] and Lehmann s work on more qualitative versions of Savage s theorem [34] Our work, too, fits within this approach since it ....

S. Tan and Judea Pearl. Qualitative decision theory. In Proceedings of the Twelfth National Conference on Artificial Intelligence, pages 928--932, 1994.

....application domains, and a synthesis mapping to do the first step in horizontal cooperation were presented. After getting the result of synthesis of solutions under uncertainty in a DES, the next step is to make a decision. Several theories to make decision under uncertainties have been proposed [1,3,6,8]. These theories are different from maximum expected utility theory, because the representations of incomplete states of knowledge are not numerical probabilities. The aims of the above decision theories are to find the best element in the frame of discernment (these decision approaches are called ....

S. W. Tan and J. Pearl, "Qualitative decision theory ", in Proceedings of the 1994 AAAI, pp. 928933.

....a new method to characterize agent s beliefs when agents gather a lot of information about the environment. The standard approach to make decisions is based on maximum expected utility theory, however, its application requires both numerical probabilities and utilities about consequences of actions. In 1994, Boutilier [ 1 ] Tan and Pearl [ 35 ] examined semantics and specification tools for qualitative decision makers, while Darwiche and Goldszmidt [ 6 ] experimented with qualitative probabilistic reasoning in diagnostics. In addition, using possibility approach, Dubois and Prade presented the ....

....when agents gather a lot of information about the environment. The standard approach to make decisions is based on maximum expected utility theory, however, its application requires both numerical probabilities and utilities about consequences of actions. In 1994, Boutilier [ 1 ] Tan and Pearl [ 35 ] examined semantics and specification tools for qualitative decision makers, while Darwiche and Goldszmidt [ 6 ] experimented with qualitative probabilistic reasoning in diagnostics. In addition, using possibility approach, Dubois and Prade presented the qualitative utility when uncertainty is ....

S. W. Tan and J. Pearl, Qualitative decision theory, in: Proceedings of AAAI, 1994, 928933.

....recently appeared in the AI community : The term qualitative decision theory refers to more than one kind of representation. Some approaches consider only all or nothing notions of utility and plausibility, for instance (Bonet and Geffner [1] others use integer valued functions (Tan and Pearl [20]) Boutilier [2] exploits preference orderings and plausibility orderings by focusing on the most plausible states. In classical decision theory under uncertainty, the preferences of the decision maker are directly expressed by means of a utility function, while a probability distribution on the ....

S. W. Tan and J. Pearl. Qualitative decision theory. In Proc. of AAAI'94, pp. 928-933, Seattle, WA, July 31-Aug. 4, 1994. Reasoning under Uncertainty 604 R. Sabbadin

....in AI applications where uncertainty is often ordinal and can be only qualitatively assessed. The same remark applies to utilities: it is often more adequate to represent preference over states simply with an ordering relation rather than with additive utilities. Recently, several authors [2] [13], 6] 7]have advocated this qualitative view of decision making and have proposed qualitative versions of decision theory, together with suitable logical languages for expressing preferences. In this paper, the possibility theory and fuzzy set based approach to the qualitative handling of ....

Sek-Wah Tan and J. Pearl, Qualitative decision theory, Proc. of the 12th National Conf. on Artificial Intelligence (AAAI'94), MIT Press, 1994, 928-933.

....decision has recently appeared in the AI community : The term qualitative decision theory refers to more than one kind of representation. Some approaches consider only all or nothing notions of utility and plausibility, for instance (Bonet and Geffner 1996) others use integervalued functions (Tan and Pearl 1994), Pearl 1993) Boutilier (1994) exploits preference orderings and plausibility orderings by focusing on the most plausible states. In (Dubois and Prade 1995) an analog of von Neumann and Morgenstern postulates, intended for rational decision under ordinal uncertainty has been proved to be ....

S. W. Tan and J. Pearl. Qualitative decision theory. InProc. 11th National Conf. on Artificial Intelligence (AAAI'94), pp. 928-933, Seattle, WA, July 31-Aug. 4, 1994.

....relation rather than with additive utilities. Recently, several authors have advocated this qualitative view of decision making and have proposed qualitative versions of decision theory, together with suitable logical languages for expressing preferences, namely, Boutilier [5] Tan and Pearl [11], Dubois and Prade [7] 8] The latter propose a qualitative utility theory based on possibility theory, where preferences and uncertainty are both qualitative. Our purpose is to extend Dubois and Prade s possibilistic framework for qualitative decision theory so as to enable multiple stage ....

Sek-Wah Tan and Judea Pearl, Qualitative decision theory, Proc. AAAI'94.

....relation rather than with additive utilities. Recently, several authors have advocated this qualitative view of decision making and have proposed qualitative versions of decision theory, together with suitable logical languages for expressing preferences, namely, Boutilier [5] Tan and Pearl [21], Dubois and Prade [13] 14] The latter propose a qualitative utility theory based on possibility theory, where preferences and uncertainty are both qualitative. Our purpose is to extend Dubois and Prade s possibilistic framework for qualitative decision theory so as to enable multiple stage ....

Sek-Wah Tan and Judea Pearl. Qualitative decision theory. Proceedings of the 12th National Conference on Artificial Intelligence (AAAI'94), MIT Press, 1994, 928-933.

....the Artificial Intelligence community : The term qualitative decision theory refers to more than one kind of representation. Some approaches consider only all or nothing notions of utility and plausibility, for instance (Bonet and Geffner [2] others use integer valued functions (Tan and Pearl [34]) Pearl [30] Boutilier [3] exploits preference orderings and plausibility orderings by focusing on the most plausible states. In (Dubois and Prade [17] a qualitative analog to von Neumann and This paper is an extended and revised version of a conference paper by the same authors [16] 2 ....

S. W. Tan and J. Pearl. Qualitative decision theory. InProc. 11th National Conf. on Artificial Intelligence (AAAI'94), pp. 928-933, Seattle, WA, July 31-Aug. 4, 1994.

....traditional approaches, see for instance [Saaty, 1980] as well as complete ordering of these properties are usually absent. Furthermore, it has been extensively argued that such approaches are impractical for the majority of decision making tasks that involve common sense knowledge and reasoning [Tan and Pearl, 1994]. In argumentation studies, subjects like priority relationships and preference orders between arguments have been mostly handled through quantitative approaches [Simari and Loui, 1992; Pinkas, 1991] In HERMES, constraints provide a qualitative way to weigh reasons for and against the selection ....

S. W. Tan and J. Pearl. Qualitative Decision Theory. In Proceedings of the AAAI'94 Conference, AAAI/MIT Press, 928-933 (1994).

....and defaults. A variant of this idea was developed by Lang (1996) who directly associates penalties with the desires (based on penalty logic (Pinkas 1995) and who does not use rankings of utility functions but utility functions themselves. More complex constructions have been discussed in (Tan Pearl 1994b; 1994a; van der Torre Weydert 2000; Lang, van der Torre, Weydert 2001b) Conflict resolution Although the constraints imposed by the I operator are rather weak, they are still too strong to represent certain types of conflicts. Consider conflicts among desires. Typically desires are allowed to be inconsistent, but once they are ....

Tan, S.-W., and Pearl, J. 1994a. Qualitative decision theory. In Proceedings of the Thirteenth National Conference on Artificial Intelligence (AAAI'94).

....to planning in AI, a planning agent is provided with a description of some state of a airs, a goal state, and charged with the task of discovering (or performing) some sequence of actions to achieve that goal. Recently several logics for goals (and the closely related desires) have been proposed [4, 3, 12, 2, 14, 13, 8, 6, 15, 16]. We are interested in a logical system that tells us which conditional goals can be derived from a set of conditional goals (the set of premises) called the goal base. In this simple setting in which we do not consider logical connectives between the conditional goals reasoning about goals is ....

S.-W. Tan and J. Pearl. Qualitative decision theory. In Proceedings of the AAAI'94, 1994.

....decision theory (Raiffa 1970) builds on the probabilistic view of uncertainty in reasoning about actions. The costs and benefits of possible outcomes of actions are weighted with their probabilities, yielding a preference ordering on the expected utility of alternative actions. However, as Tan and Pearl (1994), amongst others, have pointed out, the specification of the complete sets of probabilities and utilities required by standard decision theory make the theory impractical in complex tasks which involve common sense knowledge. This realisation has prompted work on qualitative approaches to decision ....

....is unpleasant and has significant morbidity this is preferable to loss of life, so surgery ought to be carried out. Informally we can represent this argument as in Figure 2. There are six different forms of argument in this example which has a similar scope to the examples considered by Tan and Pearl (1994). The first are those labeled e1, e5 which are standard arguments in LA. The second are value assignments v1 and v2 which represent information about what states are desirable and undesirable. The third are expected value arguments ev1 and ev2 which combine the information in standard and ....

Tan, S.-W., and Pearl, J. 1994. Qualitative decision theory. In Proceedings of the 12th National Conference on Artificial Intelligence, 928--933. Menlo Park, CA: AAAI Press/MIT Press.

....to this problem, because they allow for a systematic study and classi cation of desires by making underlying assumptions explicit. Recently several logics for desires and goals have been proposed (Doyle and Wellman, 1991; Doyle et al. 1991; Boutilier, 1994; Pearl, 1993; Tan and Pearl, 1994b; Tan c 2000 Kluwer Academic Publishers. Printed in the Netherlands. aamas.tex; 30 05 2000; 14:52; p.1 2 and Pearl, 1994a; Lang, 1996; Bell and Huang, 1997; van der Torre, 1998) to express preferences implicitly and compactly. For example, Cohen and Levesque (Cohen and Levesque, 1990) explore principles ....

....on. First, it seems appropriate to ask for fairness in the sense that the process of attaching penalties or rewards should be unbiased. That is, worlds within the same desire context should receive equal treatment, which recalls Laplace s indi erence principle for probabilities (see e.g. Neapolitan, 1990)) These considerations motivate our rst two desiderata for distinguished models of a single conditional desire. 1. Local uniformity. The agent should be indi erent with respect to any two a b worlds. Similarly for a :b worlds and :a worlds. In other words, the utility values should be constant ....

[Article contains additional citation context not shown here]

Tan, S.-W. and J. Pearl: 1994a, `Qualitative decision theory'. In: Proceedings of the Thirteenth National Conference on Articial Intelligence (AAAI'93).

....and derive e#cient inference mechanisms using them is one of the aims of the field of qualitative decision theory. Recent work from this field has attempted to address the elicitation problem by providing formal languages in which partial preference information can be conveniently expressed [9, 10, 44, 4, 1]. For example, the languages proposed by Doyle and Wellman [10] and by Tan and Pearl [44] attempt to provide a semantic to ceteris paribus (all else being equal) comparative statements. These are preferential statements concerning classes of decision consequences. While these languages have ....

....decision theory. Recent work from this field has attempted to address the elicitation problem by providing formal languages in which partial preference information can be conveniently expressed [9, 10, 44, 4, 1] For example, the languages proposed by Doyle and Wellman [10] and by Tan and Pearl [44] attempt to provide a semantic to ceteris paribus (all else being equal) comparative statements. These are preferential statements concerning classes of decision consequences. While these languages have successfully addressed a number of expressiveness issues, the inferential mechanisms available ....

[Article contains additional citation context not shown here]

S. Tan and J. Pearl. Qualitative decision theory. In Proceedings of the Twelfth National Conference on Artificial Intelligence, pages 928--932, 1994.

....the issue of representing preferences in constructing planning and automated decision making systems. The focus of this work is a formulation and interpretation of preferential comparisons, based on the notion of preference ceteris paribus (all else being equal) The work of Tan and Pearl [13, 12], which I will discuss in Section 3, is also based on a similar interpretation for conditional preference statements. In this proposal, Tan and Pearl advocate a qualitative theory for decision making that replaces probabilities and utilities with their qualitative counterparts. Boutilier [2] ....

....[2] and Bacchus and Grove [1] suggest interpretations of goals based on quite di#erent qualifiers that are much more restrictive. I will discuss these approaches in Section 4. 3 A Theory for Qualitative Decision Making In this section I discuss another line of research, pursued by Tan and Pearl [13, 12], that aims to provide a qualitative theory for decision making. The central issues in this work are formulation and interpretation of quantified conditional preferential statements, which are intended to capture all preferential information of the decision making agent. The agent s belief is ....

[Article contains additional citation context not shown here]

S. Tan and J. Pearl. Qualitative decision theory. In Proceedings of the Twelfth National Conference on Artificial Intelligence, pages 928--932, 1994.

....and derive efficient inference mechanisms using them is one of the aims of the field of qualitative decision theory. Recent work from this field has attempted to address the elicitation problem by providing formal languages in which partial preference information can be conveniently expressed [5, 6, 15, 3, 1]. For example, the languages proposed by Doyle and Wellman [6] and by Tan and Pearl [15] attempt to provide a semantic to ceteris paribus (all else being equal) comparative statements. These are preferential statements concerning classes of decision consequences. While these languages have ....

....decision theory. Recent work from this field has attempted to address the elicitation problem by providing formal languages in which partial preference information can be conveniently expressed [5, 6, 15, 3, 1] For example, the languages proposed by Doyle and Wellman [6] and by Tan and Pearl [15] attempt to provide a semantic to ceteris paribus (all else being equal) comparative statements. These are preferential statements concerning classes of decision consequences. While these languages have successfully addressed a number of expressiveness issues, the inferential mechanisms available ....

[Article contains additional citation context not shown here]

S. Tan and J. Pearl. Qualitative decision theory. In Proceedings of the Twelfth National Conference on Artificial Intelligence, pages 928--932, 1994.

....utility calculations, and the second heuristic only considers the most preferred states. The two heuristics are in an obvious way symmetric, but they have completely different consequences. The first heuristic cannot explain that people insure themselves for unlikely but grave events, see e.g. Tan and Pearl, 1994a ] and the second heuristic has the disadvantage that if the most preferred states are very unlikely, such as winning a lottery, then the requirement does not have an impact on the expected utilities and therefore not on the decisions. With the parameters the risk component of each desire can be ....

S.-W. Tan and J. Pearl. Qualitative decision theory. In Proceedings of the AAAI'94, 1994.

.... defeasible deontic logic is related to logics for qualitative decision theory (Pearl, 1993; Boutilier, 1994; van der Torre and Tan, 1999a) In particular, prima facie obligations are related to logics of desires (qualitative abstractions of utilities) which can be overridden by stronger desires (Tan and Pearl, 1994; Lang, 1996) The formal relation between obligations and desires is subject of present investigations. ....

Tan, S.-W. and Pearl, J. (1994). Qualitative decision theory.

No context found.

S.-W. Tan and J. Pearl. Qualitative decision theory. In Proceedings of the Thirteenth National Conference on Artificial Intelligence (AAAI'94), pages 928--932, 1994.

No context found.

S. Tan and J. Pearl. Qualitative decision theory. In Proceedings of the Twelfth National Conference on Artificial Intelligence, pages 928--932, 1994.

No context found.

S.-W. Tan and J. Pearl. Qualitative decision theory. In Proceedings of the Thirteenth National Conference on Artificial Intelligence (AAAI'94), pages 928--933. AAAI Press, 1994.

No context found.

S. W. Tan and J. Pearl. Qualitative decision theory. In Proceedings of the Twelfth National Conference on Arti cial Intelligence, pages 928-933, Seattle, 1994.

First 50 documents

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