Search Results

This chapter selectively reviews computational theories of analogical reasoning from Evans, Gentner, Holyoak and Thagard, Ashley, Carbonell, and Hofstadter. While these theories provide insight into the processes involved in analogical reasoning, many of them operate with a perceptual model of analogical reasoning and appear to neglect normative questions. It is argued that most of the computational theories do, at least implicitly, incorporate normative principles and that those principles need to be examined critically. In particular, the chapter takes a close look at Gentner's systematicity principle. It is alleged that systematicity per se neither produces nor explains the plausibility of analogical arguments.

This chapter provides an introduction to the paradigm of case-based reasoning, an approach to the design of knowledge-based systems which implements the idea that classification and problem-solving depend crucially on remembering similar past episodes. At the heart of any case-based system is a standard sequence of steps that has been described in detail, summarized as ‘retrieve, reuse, revise, and retain’. As a field, CBR has rapidly reached maturity, as evidenced by the number of commercial applications in use, as well as the existence of numerous development tools and methodologies. The current technological level in the field allows rapid development of help-desk-like applications, which are characterized by the simplicity of their representational structures and similarity function. However, limitations of current technology become apparent with more complex tasks.

This chapter presents a formal model that captures both case‐based and rule‐based reasoning. The model is general enough to describe Bayesian reasoning, which may be viewed as an extreme example of rule‐based reasoning. It suggests conditions under which Bayesian reasoning will give way to other modes of reasoning, and alternative conditions under which the opposite conclusion holds. It discusses how probabilistic reasoning may emerge periodically, with other modes of reasoning used between the regimes of different probabilistic models.

This chapter first adapts the framework of the model of Gilboa, Samuelson, and Schmeidler (GSS, 2013, Chapter 3) and defines case-based and rule-based reasoning. It then deals with a purely exogenous process, showing that rule-based reasoning is likely to emerge in simple states of the world, but not in all states; followed by a purely endogenous process, showing that rule-based reasoning is more likely to emerge as the asymptotic mode of reasoning than case-based reasoning. The chapter concludes that rule-based reasoning is generally not sufficient to describe the way agents think about the world—neither at the outset nor asymptotically. All other things being equal, an endogenous process is more likely to give rise to convergence to rule-based reasoning than is an exogenous one.

A goal of machine morality is not just to raise many questions but to provide a resource for further development of artificial moral agents. Chapter 9 surveys software that is currently under development for moral decision making by (ro)bots. These experiments utilize a variety of strategies including case‐based reasoning or casuistry, deontic logic, connectionism (particularism), and the prima facie duties of W. D. Ross (also related to the principles of biomedical ethics). In addition to agent approaches that focus on the reasoning of one agent, researchers are working with multi‐agent environments and with multibots. This discussion serves as a comprehensive summary of research to date directed at making (ro)bots explicit moral reasoners. These experiments range from ethical advisors in health care to strategies for ensuring that (ro)bot soldiers won't violate international conventions.

Case-based reasoning (CBR) is a problem-solving methodology that exploits prior experiences when solving new problems, retrieving relevantly similar cases and adapting them to fit new needs. This chapter describes an approach in which an introspective reasoner monitors the CBR process with the goal of adjusting the retrieval and reuse strategies of the system to improve solution quality. Among the novel aspects of this approach, compared to previous work on introspective reasoning for CBR, are that it applies a unified model for improving the two main stages of the CBR process; that a single failure may prompt multiple forms of learning; and that it performs internal tests to empirically assess the value of changes proposed by the introspective reasoner, to determine which ones should be retained.

This chapter focuses on case-based reasoning. It offers an axiomatic approach to the following problem: given a database of observations, how should different eventualities be ranked? The approach is complementary to the Bayesian approach at two levels: first, it may offer an alternative model of prediction, when the information available to the predictor is not easily translated to the language of a prior probability. Second, the approach may describe how a prior is generated. The chapter is organized as follows. Section 2 presents the formal model and the main results. Section 3 discusses the relationship to kernel methods and to maximum likelihood rankings. Section 4 presents a critical discussion of the axioms, outlining their scope of application. Finally, Section 5 briefly discusses alternative interpretations of the model, and, in particular, relates it to case‐based decision theory.

This chapter examines “casuistry,” the dominant mode of moral theology in the Roman Catholic Church that emphasizes the morally right—and the morally wrong—courses of action in a wide range of specific situations faced by Christians of all walks of life. It first offers a definition of casuistry that can account for both the manualist tradition and recent theoretical work on the topic as well as the phenomenon of case-based reasoning in the common law tradition. It then considers two sets of resources, one theoretical and one practical, that can reinvigorate interest in the Catholic casuistical tradition: the tradition theory of Alasdair MacIntyre and the Anglo-American common law tradition. It also explains how MacIntyre's tradition theory and the common law can be used to present a broader account of the context underlying the practice of good casuistry. The chapter concludes by recasting the Roman Catholic manualist tradition in an appropriate, broader context and discussing the benefits and preconditions of retrieving and reframing the Catholic casuistical tradition in a methodologically self-conscious manner.

Regulators have officially recognized what researchers have long known: human subjects review boards request different (sometimes contradictory) modifications to a given study before approval. In doing so, boards inadvertently delay research and create coordination problems, especially for multisite studies. This chapter summarizes an original ethnographic study of the meetings of US institutional review boards and introduces the concept of “local precedents” to explain why multisite research presents such a problem for boards. These ethnographic findings point to the specific shortcomings of the local-review model. To redress these shortcomings, the chapter identifies three types of review mechanisms that are configured differently from the local-review system: study networks, collegial review, and decision repositories. Several institutions are developing these alternative review mechanisms, and their experiences indicate the relative merits of each mechanism.

An ethical problem is not like a mathematics problem or most science problems that have unique solutions that are either right or wrong. Instead, ethics problems are more like design problems for which several acceptable solutions can be found. Design problems are problems of making or repairing things or processes that satisfy human desires or needs (Whitbeck 1996). The most familiar example in chemistry is design of a synthesis, an example of process design. There is usually more than one way to make a particular molecule. Deciding on which method is “best” involves a large number of considerations, including cost of materials, yield, quantity and purity of product, safety, purification methods, and reaction conditions, among others. Two different chemists might choose two different routes based on individual considerations. For example, while one route might provide a higher yield but require an expensive piece of equipment, the second route has a lower yield but can be done less expensively. The chemist who already owns the specialized equipment will probably choose the first alternative, but a colleague whose research budget is limited might accept the lower yield to save money. In a second kind of synthesis design problem, the end use is known, but several molecules or materials might actually accomplish this goal. Drug design is a good example. A chemist might take on (or be assigned) the task of developing a compound that controls blood pressure by blocking an enzyme that constricts blood vessels. A number of compounds might work, and the “best” solution to the problem will depend on factors such as ease of synthesis and purification, cost, medical side effects, and safety and environmental considerations involved in the manufacture of the drug. In general, the design’s success depends on whether it achieves the desired end within the imposed criteria and constraints. There is a close analogy between design problems and real-life ethical problems. In an ethical problem, a chemist or chemistry student must devise possible courses of action, evaluate them, and then decide what to do.