J. Doyle. Constructive belief and rational representation. Computational Intelligence, 5:1-11, 1989.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....and Schmeidler [26] examines case based decision making, which should be of interest to researchers in qualitative decision theory. 22 Although researchers in arti cial intelligence (and to some extent statisticians) have examined and used various qualitative notions of knowledge, belief (e.g. [12, 29, 8]) and preference (e.g. 15] most e ort has been on de ning these concepts and understanding the process of (qualitative) belief revision following new information. More recently, ideas on how qualitative notions of belief and preference can be combined have been examined [44, 5] However, ....

J. Doyle. Constructive belief and rational representation. Computational Intelligence, 5:1-11, 1989.

....and Schmeidler [26] examines case based decision making, which should be of interest to researchers in qualitative decision theory. 22 Although researchers in artificial intelligence (and to some extent statisticians) have examined and used various qualitative notions of knowledge, belief (e.g. [12, 29, 8]) and preference (e.g. 15] most effort has been on defining these concepts and understanding the process of (qualitative) belief revision following new information. More recently, ideas on how qualitative notions of belief and preference can be combined have been examined [44, 5] However, ....

J. Doyle. Constructive belief and rational representation. Computational Intelligence, 5:1--11, 1989.

....entailments and non monotonic entailments, as well as deductive entailments. A representer who conveys knowledge as p and if p then q also conveys as knowledge q , implicitly. This is what is meant by using both sentences p and if p then q in a propositional language, pc. Both Jon Doyle (1989) and I (1991) have recently written at length about this view of implicit knowledge and its alternatives. Philosophers have called non ampliative all inferences that merely 2 For a discussion of Aristotle and dialectics, see Hogan. For a discussion of Keynes, see Carabelli. For Aquinas, see ....

....an audience as possible, not due to a lack of perception. Pollock s record of work in conditional logics gives him a stake in substantive approaches. His history of leadership in epistemology asks that he see defeasibility arising from further experience as well as from further computation. In AI, Doyle (1989) has been making about representation and belief many of the same claims made here. Among them, Doyle agrees that belief is constructed from manifest representation, and that sometimes choice is involved. Doyle also sees belief construction as an activity, one step in the more general process ....

Doyle, J. "Constructive belief and rational representation," Computational Intelligence 5, 1989.

....in this paper, and this work will be discussed in the next section. However, the most closely related lines of research can be found within economics, game theory, and decision theory, areas whose relevance to AI research has been pointed out by many researchers, most notably, Jon Doyle (e.g. [18, 19]) In particular, three topics of research are directly relevant to our work: work on subjective probability and choice theory, work on qualitative decision making, and work on revealed preference. In what follows, we briefly describe these fields and compare them with our own effort. 7.1 ....

J. Doyle. Constructive belief and rational representation. Computational Intelligence, 5:1--11, 1989.

....to decide whether some default is correct or not. When we do take the semantics seriously it is not so obvious that the default statements say what we actually want to say. 1. 1 Defaults and utilities What one is prepared to say yes to depends on both utility and probabilistic information (Doyle [ 1989 ] argues this most strongly; see section 6.1) If one is playing a game like trivial pursuit (where there is no penalty for saying something wrong over the penalty for saying nothing) it is better to have a wild guess at something than to say nothing. If one is in court acting as an expert ....

....with certain properties) does affect the defaults we make. In this paper we consider a formulation of defaults that takes probability and utility into consideration. 1. 2 The Proposal Other people have observed that utilities have something to do with default reasoning [ Shoham, 1987; Loui, 1990; Doyle, 1989; Kadie, 1988 ] In this paper we take this relationship seriously and treat defaults as decision theory summaries. A default e A a means that a is the best decision out of those decisions in A if all you know is e. Note that the conclusion of a default is an action, and not a proposition as in ....

[Article contains additional citation context not shown here]

J. Doyle. Constructive belief and rational representation. Computational Intelligence, 5(1):1--11, February 1989.

....locales, but the current implementation of AMORD conducts all of its reasoning in a single locale. The RMS must also relabel nodes from in to out or vice versa as necessitated by the reasoning being performed. For more discussion of reason maintenance and nonmonotonic reasoning see [6] [7], 8] 9] 11] and [10] The RMS uses the Reasoning Economy (RECON) to determine which relabelings to perform when. 6.1.3 RECON As mentioned in section 3.4, the Reasoning Economy is built on top of Wellman s WALRAS market oriented programming environment. RECON creates a good for each ....

Jon Doyle. Constructive belief and rational representation. Computational Intelligence, 5(1):1--11, February 1989.

.... on default and defeasible reasoning (e.g. McCarthy, 1980 ] In particular, several works use expected utility consideration in evaluation of heuristic rules (e.g. Langlotz et al. 1986 ] More recently, decision theoretic foundations for defaults were advocated by Shoham (1987) and Doyle (1989). Doyle provides a formal analysis of Pascal s wager and shows how an assumption (the existence of God) can be justified in terms of utility. Finally, Poole (1992) examined a concrete notion of defaults that are grounded in terms of decision theory. Unlike previous works (with the exception of ....

J. Doyle. Constructive belief and rational representation. Computational Intelligence, 5:1-- 11, 1989.

....can choose. Game theory allows for multiple moves by different agents and by nature. ffl Utility and values play an integral part in decision and game theory. They are not part of nonmonotonic formalisms, although it has been admitted that values do play a part in what assumptions should be made [45, 8]. Utilities have not been explicit, and maybe they need to be so that they can be reasoned about and not compiled into a set of assumptions. ffl What information is available to agents when making a decision is also important (as we do not always have perfect information ) This plays an ....

J. Doyle. Constructive belief and rational representation. Computational Intelligence, 5(1):1--11, February 1989.

....best performance. Making control decisions rationally raises the problem of infinite regress, since trying to control the cost of making rational control decisions by means of additional rational control decisions creates a tower of deliberations, each one concerned with the level below (as in (Doyle, 1980)) 11 Thus striking a balance between control and reasoning computations means taking effort expended at all these levels into account. In practice, the deliberative information available at higher levels but unavailable at lower ones vanishes as one ascends the reflective tower, and most ....

....We need to investigate how to integrate these theories in useful ways that recognize that meaning, possibility, utility, and probability must all be evaluated with respect to changing purposes and circumstances. 31 Doyle Acknowledgments This paper is an extended version of an invited talk (Doyle, 1990) presented at AAAI90. I thank Ramesh Patil, Peter Szolovits, and Michael Wellman for reading drafts, Rich Thomason for lending me some of his notes, Tom Dean, Othar Hansson, Eric Horvitz, Barton Lipman, Andrew Mayer, Stuart Russell, Joseph Schatz, and David Smith for valuable discussions, and the ....

Doyle, J. 1989a. Constructive belief and rational representation. Computational Intelligence, 5(1):1--11.

....the value of the services provided to the user or to the larger reasoning system making use of the knowledge base. Instead, a better view is that the purpose of representational services is to provide the rational or optimal conclusions rather than the logically sound conclusions (see also Doyle [9]) and that representation systems should be rational agents using the user s knowledge and purposes to cooperate with the user to accomplish those purposes. We present the detailed arguments for each of the criticisms and alternative proposals after first setting out the fundamentals of taxonomic ....

J. Doyle. Constructive belief and rational representation. Computational Intelligence, 5(1):1--11, February 1989.

.... agents, one does better to frame states in terms of two sets of state components; a set of manifest components explicitly contained or represented in the state, and a set of constructive components implicitly stored or represented in the state and computed or derived from the manifest components (Doyle, 1989). Accordingly, we define a constructive framing to be a multiattribute framing OE : S Sm Theta S c interpreting each state as a manifest state in a set Sm and a constructive state in a set S c , and write OE m : S Sm and OE c : S S c to mean the corresponding projections of this mapping ....

Doyle, J. (1989). Constructive belief and rational representation. Computational Intelligence, 5 (1), 1--11.

....as the strongest argument for acceptance. Pragmatic efficiency strikes me as too narrow a criterion unless we interpret efficiency in its most general sense as economic rationality taking into account the impact of beliefs on the agent s situation as well as computational costs. My papers [1, 2, 3, 4, 5] provide more details on my view of rational belief and inference, generally taking a coordinate free approach in which one may consider alternative ways of reading beliefs out of mental states. This permits, for example, interpreting specific beliefs and inference in any of the ways discussed ....

J. Doyle. Constructive belief and rational representation. Computational Intelligence, 5(1):1--11, Feb. 1989.

....aggregation. We examine how irrationalities in preference aggregation lead to violations of Gardenfors s rationality axioms. Section 6 expands the theory further to cover rational revision of preferences. This final expansion, which is related to our earlier theory of constructive attitudes [11], introduces further ways in which rational belief revision can diverge from Gardenfors s notion. 2 Revising belief states Gardenfors s formalization of belief revision may be summarized as follows, using an adaptation of Nebel s notation. We suppose that L is a propositional language over the ....

....always exist. It is not clear that this assumption is necessary, or that the definitions here are the best possible. How to deal with the case in which no alternatives are maximally preferred is an interesting question for future study. It would appear that the notion of rational representation [10, 11], in which consistent subsets of inconsistent preferences are chosen to temporarily represent the inconsistent sets, should play a role here. The generality exhibited in the definitions of rational contraction and revision goes against the usual presupposition of epistemologists that knowing more ....

[Article contains additional citation context not shown here]

J. Doyle. Constructive belief and rational representation. Computational Intelligence, 5(1):1--11, Feb. 1989.

....inference is more than just an interpretation: it provides a justification for the formal structures of the various nonmonotonic logics. The original theories provided precise formal concepts, but motivated explanations of why these concepts were interesting appeared only later, when Doyle [6, 9], Shoham [55] and others [26, 28] justified default rules by an appeal to decision theoretic rationality, saying that an agent should adopt a default conclusion or default rule if the expected inferential utility of holding it exceeds that of not holding it. Default rules and other ....

Doyle, J. Constructive belief and rational representation. Computational Intelligence 5 (1989) 1--11.

....cases nonlogical restrictions might be naturally made part of the fixed constitution, as global conditions on states which refine the constitutive logic. One example might be the rationality of the set of constructive attitudes contained in a state with respect to the state s manifest attitudes [ Doyle, 1989 ] But many sorts of nonlogical restrictions, such as individual defaults or reason maintenance justifications, are naturally viewed as locally applicable laws of thought or constitutive intentions. Constitutive intentions are intentions strictly about the agent s cognitive structure as opposed ....

Jon Doyle. Constructive belief and rational representation. Computational Intelligence, 5(1):1--11, February 1989.

No context found.

Doyle, J. "Constructive belief and rational representation," Computational Intelligence 5, 1989.

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