Rationality and Intelligence

The long-term goal of our field is the creation and understanding of intelligence. Productive research in AI, both practical and theoretical, benefits from a notion of intelligence that is precise enough to allow the cumulative development of robust systems and general results. This paper outlines a gradual evolution in our formal conception of intelligence that brings it closer to our informal conception and simultaneously reduces the gap between theory and practice. 1 Artificial Intelligence AI is a field in which the ultimate goal has often been somewhat ill-defined and subject to dispute. Some researchers aim to emulate human cognition, others aim at the creation of intelligence without concern for human characteristics, and still others aim to create useful artifacts without concern for abstract notions of intelligence. This variety is not necessarily a bad thing, since each approach uncovers new ideas and provides fertilization to the others. But one can argue that, since philosophers abhor a definitional vacuum, many of the damaging and ill-informed debates about the feasibility of AI have been about definitions of AI to which we as AI researchers do not subscribe. My own motivation for studying AI is to create and understand intelligence as a general property of systems, rather than as a specific attribute of humans. I believe this to be an appropriate goal for the field as a whole, and it certainly includes the creation of useful artifacts—both as a spin-off and as a focus and driving force for technological development. The difficulty with this “creation of intelligence” view, however, is that it presupposes that we have some productive notion of what intelligence is. Cognitive scientists can say “Look, my model correctly predicted this experimental observation of human cognition,” and artifact developers can say “Look, my system is saving lives/megabucks,” but few of us are happy with papers saying “Look, my system is intelligent.” This difficulty is compounded further by the need for theoretical scaffolding to allow us to design complex systems with confidence and to build on the results of others. “Intelligent” must be given a definition that can be related directly to the system’s input, structure, and output.1 Such a definition must also be general. Otherwise, AI subIn this paper, I shall outline the development of such definitions over the history of AI and related disciplines.2 I shall examine each definition as a predicate P that can be applied, supposedly, to characterize systems that are intelligent. For each P, I shall discuss whether the statement “Look, my system is P” is interesting and at least sometimes true, and the sort of research and technological development to which the study of P-systems leads. I shall begin with the idea that intelligence is strongly related to the capacity for successful behaviour—the so-called “agent-based” view of AI. The candidates for formal definitions of intelligence are as follows: P1: Perfect rationality, or the capacity to generate maximally successful behaviour given the available information. P2: Calculative rationality, or the in-principle capacity to compute the perfectly rational decision given the initially available information. P3: Metalevel rationality,or the capacity to select the optimal combination of computation-sequence-plus-action, under the constraint that the action must be selected by the computation. P4: Bounded optimality, or the capacity to generate maximally successful behaviour given the available information and computational resources. All four definitions will be fleshed out in detail, and I will describe some results that have been obtained so far along these lines. Then I will describe ongoing and future work under the headings of calculative rationality and bounded optimality. I shall be arguing that, of these candidates, bounded optimality comes closest to meeting the needs of AI research. There is always a danger, in this sort of claim, that its acceptance can lead to “premature mathematization,” a condition characterized by increasingly technical results that have increasingly little to do with the original problem—in the case of AI, the problem of creating intelligence. Is research on bounded optimality a suitable stand-in for research on intelligence? I hope to show that P4, bounded optimality, is closer than P1 through P3 because it is a real problem with real and desirable solutions, and also because it satisfies some sides into a smorgasbord of fields—intelligence as chess playing, intelligence as vehicle control, intelligence as medical diagnosis. In doing so I shall draw heavily on previous work with Eric Wefald [Russell and Wefald, 1991a] and Devika Subramanian [Russell and Subramanian, 1995]. The latter paper contains a much more rigorous analysis of the concepts presented here. essential intuitions about the nature of intelligence. Some important questions about intelligence can only be formulated and answered within the framework of bounded optimality or some relative thereof. Only time will tell, however, whether bounded optimality research, perhaps with additional refinements, can generate enough theoretical scaffolding to support significant practical progress in AI.