|Norvig I 611
Utility/AI research/Norvig/Russell: The principle of maximum expected utility (MEU) says that a rational agent should choose the action that maximizes the agent’s expected utility (…).In a sense, the MEU principle could be seen as defining all of AI. All an intelligent agent has to do is calculate the various quantities, maximize utility over its actions, and away it goes. But this does not mean that the AI problem is solved by the definition!
The MEU principle formalizes the general notion that the agent should “do the right thing,” but goes only a small distance toward a full operationalization of that advice. Estimating the state of the world requires perception, learning, knowledge representation, and inference. Computing P(RESULT(a) | a, e) requires a complete causal model of the world and, (…), NP-hard inference in (very large) >Bayesian networks. >Learning/AI research, >Knowledge representation/AI research, >Inference/AI research, >Decisions/AI research.
Justification of the MEU principle: If an agent acts so as to maximize a utility function that correctly reflects the performance measure, then the agent will achieve the highest possible performance score (averaged over all the possible environments). >Utility theory/Norvig.
Norvig I 615
Def Utility/Norvig/Russell: Utility is a function that maps from lotteries to real numbers. >Utility theory/Norvig, >Rationality/AI research, >Certainty effect/Kahneman/Tversky, >Ambiguity/Kahneman/Tversky.
Norvig I 637
The derivation of numerical utilities from preferences was first carried out by Ramsey (1931)(1); the axioms for preference in the present text are closer in form to those rediscovered in Theory of Games and Economic Behavior (von Neumann and Morgenstern, 1944)(2). A good presentation of these axioms, in the course of a discussion on risk preference, is given by Howard (1977)(3). Ramsey had derived subjective probabilities (not just utilities) from an agent’s preferences; Savage (1954)(4) and Jeffrey (1983)(5) carry out more recent constructions of this kind. Von Winterfeldt and Edwards (1986)(6) provide a modern perspective on decision analysis and its relationship to human preference structures. The micromort utility measure is discussed by Howard (1989)(7).
1. Ramsey, F. P. (1931). Truth and probability. In Braithwaite, R. B. (Ed.), The Foundations of Mathematics and Other Logical Essays. Harcourt Brace Jovanovich.
2. von Neumann, J. and Morgenstern, O. (1944). Theory of Games and Economic Behavior (first edition).
Princeton University Press.
3. Howard, R. A. (1977). Risk preference. In Howard, R. A. and Matheson, J. E. (Eds.), Readings in Decision
Analysis, pp. 429–465. Decision Analysis Group, SRI International.
4. Savage, L. J. (1954). The Foundations of Statistics. Wiley.
5. Jeffrey, R. C. (1983). The Logic of Decision (2nd edition). University of Chicago Press.
6. von Winterfeldt, D. and Edwards,W. (1986). Decision Analysis and Behavioral Research. Cambridge
7. Howard, R. A. (1989). Microrisks for medical decision analysis. Int. J. Technology Assessment in
Health Care, 5, 357–370._____________Explanation of symbols: Roman numerals indicate the source, arabic numerals indicate the page number. The corresponding books are indicated on the right hand side. ((s)…): Comment by the sender of the contribution. The note [Author1]Vs[Author2] or [Author]Vs[term] is an addition from the Dictionary of Arguments. If a German edition is specified, the page numbers refer to this edition.
Stuart J. Russell
Artificial Intelligence: A Modern Approach Upper Saddle River, NJ 2010