Catálogo de publicaciones - libros

Cognitive psychologist Patricia Cheng suggests that erroneous causal inference is perhaps too often incorrectly attributed to problems with the of inference rather than the on which the inference is carried out. In this paper, we discuss the role of incomplete data in making faulty inferences and where those problems arise. We focus on one of two potential problems in the data we call and and address a generalization of the causal knowledge in the hope of detecting independencies hidden inside variables, causing the system to behave less than adequately.

The interpretation of the data can be more representative of the problem domain by examining subsets of values for variables in the data. We show how to do this with a generalized form of statistical independence that can resolve relevance problems in the causal model. The most interesting finding is how the examination of contexts can formalize the paradoxical statements in Simpson’s paradox and how a simple detection method can eliminate the problem.

Causal independence models offer a high level starting point for the design of Bayesian networks but are not maximally exploited as their behaviour is often unclear. One approach is to employ qualitative probabilistic network theory in order to derive a qualitative characterisation of causal independence models. In this paper we exploit polynomial forms of Boolean functions to systematically analyse causal independence models, giving rise to the notion of a polynomial causal independence model. The advantage of the approach is that it allows understanding qualitative probabilistic behaviour in terms of algebraic structure.

The aim of this paper is twofold: The first objective is to study the degree of dominance of fuzzy choice functions, a notion that generalizes Banerjee’s concept of dominance. The second objective is to use the degree of dominance as a tool for solving multicriteria decision making problems. These types of problems describe concrete economic situations where partial information or human subjectivity appears. The mathematical modelling is done by formulating fuzzy choice problems where criteria are represented by fuzzy available sets of alternatives.

The paper presents a first tentative work that investigates the interest and the questions raised by the introduction of argumentation capabilities in multiple criteria decision-making. Emphasizing the positive and the negative aspects of possible choices, by means of arguments in favor or against them is valuable to the user of a decision-support system. In agreement with the symbolic character of arguments, the proposed approach remains qualitative in nature and uses a bipolar scale for the assessment of criteria. The paper formalises a multicriteria decision problem within a logical argumentation system. An illustrative example is provided. Various decision principles are considered, whose psychological validity is assessed by an experimental study.

In this paper we introduce and study a nonmonotonic logic to reason about various kinds of preferences. We introduce preference types to choose among these kinds of preferences, based on an agent interpretation. We study ways to calculate “distinguished” preference orders from preferences, and show when these distinguished preference orders are unique. We define algorithms to calculate the distinguished preference orders.

This paper proposes an approach to representing preferences about multifactorial ratings. Instead of defining a scale of values and aggregation operations, we propose to express rationality conditions and other generic properties, as well as preferences between specific instances, by means of constraints restricting a complete pre-ordering among tuples of values. The derivation of a single complete pre-order is based on possibility theory, using the minimal specificity principle. Some hints for revising a given preference ordering when new constraints are required, are given. This approach looks powerful enough to capture many aggregation modes, even some violating co-monotonic independence.

Decisions can be assessed by sets of positive and negative arguments — the problem is then to compare these sets. Studies in psychology have shown that the scale of evaluation of decisions should then be considered as bipolar. The second characteristic of the problem we are interested in is the qualitative nature of the decision process — decisions are often made on the basis of an ordinal ranking of the arguments rather than on a genuine numerical evaluation of their degrees of attractiveness or rejection. In this paper, we present and axiomatically characterize two methods based on possibilistic order of magnitude reasoning that are capable of handling positive and negative affects. They are extensions of the maximin and maximax criteria to the bipolar case. More decisive rules are also proposed, capturing both the Pareto principle and the idea of order of magnitude reasoning.

This paper is centered on the family of Dung’s finite argumentation frameworks when the attacks relation is symmetric (and nonempty and irreflexive). We show that while this family does not contain any well-founded framework, every element of it is both coherent and relatively grounded. Then we focus on the acceptability problems for the various semantics introduced by Dung, yet generalized to sets of arguments. We show that only two distinct forms of acceptability are possible when the considered frameworks are symmetric. Those forms of acceptability are quite simple, but tractable; this contrasts with the general case for which all the forms of acceptability are intractable (except for the ones based on grounded or naive extensions).

Analyzing argumentation semantics with respect to the notion of skepticism is an important issue for developing general and well-founded comparisons among existing approaches. In this paper, we show that the notion of skepticism plays also a significant role in order to better understand the behavior of a specific semantics in different situations. Building on an articulated classification of argument justification states into seven distinct classes and on the definition of a weak and a strong version of skepticism relation, we define the property of skepticism adequacy of an argumentation semantics, which basically consists in requiring a lesser commitment when transforming a unidirectional attack into a mutual one. We then verify the skepticism adequacy of some literature proposals and obtain the rather surprising result that some semantics fail to satisfy this basic property.

In order to give full support for differential diagnosis of dementia in medical practice, one single clinical guideline is not sufficient. A synthesis guideline has been formalized using core features from selected clinical guidelines for the purpose of providing decision support for clinicians in clinical practice. This guideline is sufficient for typical cases in the domain, but in order to give support in atypical cases additional clinical guidelines are needed which are pervaded with more uncertainty. In order to investigate the applicability of a probabilistic formalism language for the formalization of these guidelines, a case study was made using the qualitative probabilistic reasoning approach developed in [1]. The case study is placed in context of a foundational view of transformations between logics. The clinical decision-making motivation and utility for this transformation will be given together with some formal indications concerning this transformation.