Abductive reasoning

About: Abductive reasoning is a research topic. Over the lifetime, 1917 publications have been published within this topic receiving 44645 citations. The topic is also known as: abduction & abductive inference.

The making of an expert detective : Thinking and deciding in criminal investigations

Abstract: Fahsing, I.A. (2016). The Making of an Expert Detective: Thinking and Deciding in Criminal Investigations. Department of Psychology, University of Gothenburg, Sweden. Drawing on theoretical frameworks developed in social and cognitive psychology, this thesis examines the degree to which individual and systemic factors may compensate for inherent biases in criminal detectives’ judgments and decision-making. Study I – an interview study – explored criminal detectives’ views of critical factors related to decision making in homicide investigations. Experienced homicide investigators in Norway (n = 15) and the UK (n = 20) were asked to identify decisional ‘tipping point’– decisions that could change detectives’ mind-set from suspect identification to suspect verification together with situational and individual factors relating to these decisions. In a content analysis, two types of decision were identified as typical and potentially critical tipping-points: (1) decisions to point-out, arrest, or charge a suspect, and (2) decisions concerning main strategies and lines of inquiry in the case. Moreover, 10 individual factors (e.g. experience) and 14 situational factors (e.g. who is the victim) were reported as related to the likelihood of mind-set shifts, most of which correspond well with previous decisionmaking research. Study II, using a quasi-experimental design, compared the quality of investigative decisions made by experienced detectives and novice police officers in two countries with markedly different models for the development of investigative expertise (England and Norway). In England, accredited homicide detectives vastly outperformed novice police officers in the number of adequate investigative hypotheses and actions reported. In Norway, however, bachelor educated police novices did marginally better than highly experienced homicide detectives. Adopting a similar design and the same stimulus material, Study III asked if a general test of cognitive abilities used in the selection process at the Norwegian Police University College could predict police students’ ability to generate investigative hypotheses. The findings did not support such a notion and this is somewhat in line with the available knowledge in the area showing that cognitive ability tests have low predictability for applied reasoning tasks. Taken together, this thesis suggests that investigative judgments are highly susceptible to the individual characteristics and biases of the detective. The results indicate that detective-expertise might act as a viable safeguard against biased decision-making, but length of experience alone does not predict sound judgments or decisions in critical stages of criminal investigations. Education and training is a solid foundation for the making of an expert detective. Nevertheless all participants’ researched across the two experiments were biased towards crime and guilt assumptive hypotheses. Hence, true abductive reasoning (i.e. to identify all competing explanations) and the presumption of innocence is hard to operationalise even for expert detectives with extensive

On Abduction in Design

Abstract: The mechanism of design reasoning from function to form is addressed by examining the possibility of explaining it as abduction. We propose a new interpretation to some definitions of innovative abduction, to show first that the concept, idea, as the basis for solution must be present in the inference, and second, that the reasoning from function to form is best modeled as a two-step inference, both of the innovative abduction pattern. This double-abductive reasoning is shown also to be the main form of reasoning in the empirically-derived “parameter analysis” method of conceptual design. Finally, the introduction of abduction into design theory is critically assessed, and in so doing, topics for future research are suggested.

Abductive reasoning and learning

Abstract: On the Relation between Abduction and Inductive Learning P.A. Flach, A.C. Kakas. Part I: Logical Approaches. AI Approaches to Abduction G. Paul. Abduction in Labelled Deductive Systems D.M. Gabbay. Logical Characterisations of Inductive Learning P.A. Flach. Abduction in Machine Learning F. Bergadano, V. Cutello, D. Gunetti. Part II: Numerical Approaches. An Overview of Ordinal and Numerical Approaches to Causal Diagnostic Problem Solving D. Dubois, H. Prade. Abductive Inference with Probabilistic Networks C. Borgelt, R. Kruse. Learning from Data: Possibilistic Graphical Models J. Gebhardt. Independence in Uncertainty Theories and its Applications to Learning Belief Networks L.M. de Campos, J.F. Huete, S. Moral. Index.

Probabilistic text understanding

Abstract: We discuss a new framework for text understanding. Three major design decisions characterize this approach. First, we take the problem of text understanding to be a particular case of the general problem of abductive inference. Second, we use probability theory to handle the uncertainty which arises in this abductive inference process. Finally, all aspects of natural language processing are treated in the same framework, allowing us to integrate syntactic, semantic and pragmatic constraints. In order to apply probability theory to this problem, we have developed a probabilistic model of text understanding. To make it practical to use this model, we have devised a way of incrementally constructing and evaluating belief networks. We have written a program,wimp3, to experiment with this framework. To evaluate this program, we have developed a simple ‘single-blind’ testing method.