Haut PDF Perspectives on belief and change

Perspectives on belief and change

Perspectives on belief and change

2.3. A new approach: internal versus external perspectives of the world 19 correctly represented by the multi-agent possible world (M, w) then for her w is the (only) ac- tual world. In fact the other possible worlds of a multi-agent possible world are just present for technical reasons: they express the other agents’ beliefs (in world w). One could get rid of the condition that a multi-agent possible world (M, w) is generated by w but the worlds which do not belong to the submodel generated by w would have neither philosophical nor technical motivation. Besides, for the same reason that Φ is finite, a multi-agent possible world is also assumed to be finite. Finally, notice that we assume that accessibility rela- tions are serial, transitive and euclidean. This means that the agents’ beliefs are consistent and that agents know what they believe and disbelieve (see Section 2.2.2). These seem to be very natural constraints to impose on the notion of belief. Intuitively, this notion of be- lief corresponds for example to the kind of belief in a theorem that you have after having proved this theorem and checked the proof several times. In the literature, this notion of be- lief corresponds to Lenzen’s notion of conviction [Lenzen, 1978] or to Gärdenfors’ notion of acceptance [Gärdenfors, 1988] or to Voorbraak’s notion of rational introspective belief [Voor- braak, 1993]. In fact, in all the agent theories the notion of belief satisfies these constraints: in Cohen and Levesque’s theory of intention [Cohen and Levesque, 1990] or in Rao and Georgeff BDI architecture [Georgeff and Rao, 1991] [Rao and Georgeff, 1991] or in Meyer et. al. KARO architecture [van Linder et al., 1998] [Meyer et al., 2001] or in Wooldridge BDI logic LORA [Wooldridge, 2000]. However, one should note that all these agent theories follow the external approach and thus use standard epistemic models (defined in Definition 2.2.1) to represent the situation. This is of course at odds with their intention to implement their theories in machines. 3
En savoir plus

179 En savoir plus

Belief revision, minimal change and relaxation: A general framework based on satisfaction systems, and applications to description logics

Belief revision, minimal change and relaxation: A general framework based on satisfaction systems, and applications to description logics

belief evolve with newly acquired knowledge. Belief expansion consists in adding new knowledge without checking consistency, while both contraction and revision consist in consistently removing and adding new knowledge, respectively. When knowledge bases are logical theories, i.e. a set of sentences in a given logic, these changes are governed by a set of postulates proposed for the first time by Alchour- ròn, Gardenfors and Makinson [1], and since known as the AGM theory. Although defined in the abstract framework of logics given by Tarski [38] (so called Tarskian logics), postulates of the AGM theory make strong assumptions on the considered logics. Indeed, in [1] the considered logics have to be closed under the standard propositional connectives in {∧, ∨, ¬, ⇒}, to be compact (i.e. property entailment depends on a finite set of axioms), and to satisfy the deduction theorem (i.e. en- tailment and implication are equivalent). While compactness is a standard property of logics, to be closed under the standard propositional connectives is more ques- tionable. Indeed, many non-classical logics such as description logics, equational logic or Horn clause logic, widely used for various modern applications in com- puting science, do not satisfy such a constraint. Recently, in many works, belief change has been studied in such non-classical logics [12,17,33,34]. In this direc- tion, we can cite Ribeiro & al.’s work in [34] that studies contraction at the abstract level of Tarskian logics, and the recent work in [40] on the extension of AGM contraction to arbitrary logics. The adaptation of AGM postulates for revision for non-classical logics has been studied but only for specific logics, mainly description logics [16,17,27,28,30,32,39] and Horn logics [11,41]. The reason is that revision can be abstractly defined in terms of expansion and retraction following the Levi identity, but this requires the use of negation, which rules out some non-classical logics [33].
En savoir plus

37 En savoir plus

Non-conscious Effects of Marketing Communication and Implicit Attitude Change: State of Research and New Perspectives

Non-conscious Effects of Marketing Communication and Implicit Attitude Change: State of Research and New Perspectives

After a critical assessment of methodologies currently used in this field of research, we propose new perspectives aimed at enhancing the theoretical, ecological and predictive validity of methodologies and dependent variable measures. An enhanced methodological validity is of a two-fold major interest. First, a scientific interest as far as theoretical advances in the non-conscious influences closely depends on the development of new valid methods. Secondly, at a practical level, because the measures of effects currently used by advertising agencies in copy-testing, post-test or tracking underestimate the effects of marketing communication campaigns. Greenwald and Banaji [7] defined implicit attitude as “introspectively unidentified (or inaccurately identified) traces of past experience that mediate favorable or unfavorable feeling, thought or action toward social objects”. Within a marketing communication context, we are interested in the antecedents on the one hand, and in the formation and change processes on the other hand, and finally, in the consequences of the implicit attitude towards a brand. Thus, the article focuses more extensively on non-conscious processes operating in the persuasion process. According to Dijksterhuis et al. [8] and Cheesman and Merikle [9], we start from an initial definition of the “non-conscious”, which will be examined and refined over the article. Psychological activities (stimuli processing, processes…) are non-conscious when an individual is unable to intentionally produce a symbolic answer (e.g. a verbal reply) relating to these activities. An important part of the processes which underlie the implicit attitude change are automatic and therefore non conscious [10]. We limit the scope of this article to the three main types of marketing communication whose non-conscious influences have been specifically examined: advertising in the media and on the Internet, product placement in movies and sponsorship in the media.
En savoir plus

15 En savoir plus

The rationality of belief change and the unexpected effects of a conflict of values

The rationality of belief change and the unexpected effects of a conflict of values

All this being said, to what degree can examination of these two belief mechanisms help to clarify and explain the processes of abandonment of belief and disaffiliation? Although a factual contradiction causes a minimal revision of belief, its limited effects on the belief system make it more difficult to reach a complete understanding of the underlying process. In contrast, the spreading effects of an axiological contradiction, which can reach to the very core of the belief system, help clarify the issues surrounding belief change. Nonetheless, as mentioned in the second example given, this type of contradiction does not lead to disaffiliation. Therefore, we must consider another explanation—the mnesic trace. “A mnesic trace concerns the process by which an individual encodes significant events in their autobiographical memory (Brewer, 1986; Conway, Pleydell-Pearce, 2000). This process is highly impacted by emotional responses (Brown and Kulik, 1977; Oishi et al., 2007; Buchanan, 2007; Holland and Kensinger, 2010). When an intense period of doubt follows recognition of a contradiction, the follower experiences so strong a series of emotions that the past event, facts, and mental states (Tulving, 1985) are memorized with the highest degree of accuracy (Talarico, LaBar, and Rubin, 2004). These mechanisms of memory—and particularly the mnesic trace—can help our understanding of the major impact axiological contradiction can have on the follower's beliefs.
En savoir plus

20 En savoir plus

Belief revision, minimal change and relaxation: A general framework based on satisfaction systems, and applications to description logics

Belief revision, minimal change and relaxation: A general framework based on satisfaction systems, and applications to description logics

LTCI, CNRS, Télécom ParisTech, Université Paris-Saclay, Paris, France, isabelle.bloch@telecom-paristech.fr Abstract Belief revision of knowledge bases represented by a set of sentences in a given logic has been extensively studied but for specific logics, mainly propositional, and also recently Horn and description logics. Here, we propose to generalize this operation from a model- theoretic point of view, by defining revision in an abstract model theory known under the name of satisfaction systems. In this framework, we generalize to any satisfaction systems the characterization of the well known AGM postulates given by Katsuno and Mendelzon for propositional logic in terms of minimal change among interpretations. Moreover, we study how to define revision, satisfying the AGM postulates, from relaxation notions that have been first introduced in description logics to define dissimilarity measures between concepts, and the consequence of which is to relax the set of models of the old belief un- til it becomes consistent with the new pieces of knowledge. We show how the proposed general framework can be instantiated in different logics such as propositional, first-order, description and Horn logics. In particular for description logics, we introduce several con- crete relaxation operators tailored for the description logic ALC and its fragments EL and ELU, discuss their properties and provide some illustrative examples.
En savoir plus

38 En savoir plus

Scientific truth or debate: On the link between perceived scientific consensus and belief in anthropogenic climate change

Scientific truth or debate: On the link between perceived scientific consensus and belief in anthropogenic climate change

From another research perspective, perceived scientific consensus as a belief is considered to follow - not shape - pre-existing climate change beliefs and worldviews. Kahan et al. (2011) take this perspective and claim that the perception of scientific consensus (or of issues such as how credible, trustworthy and/or competent scientists are) is itself determined by pre-existing individual beliefs shaped by culture. The theory of the cultural cognition of risk proposes that scientific information and claims of consensus are overall more readily recalled when they fit the individuals’ own beliefs on the matter, and “[help] to explain public disagreement about the significance of empirical evidence generally” (Kahan et al., 2011, p. 148). Kahan and colleagues have demonstrated that respondents holding hierarchical- individualistic worldviews perceive significantly lower scientific consensus on anthropogenic climate change than those holding egalitarian-communitarian worldviews (Kahan et al., 2011, Study 1). Moreover, higher levels of trust and competence are attributed to experts whose positions are aligned (versus non-aligned) with those of the participants (Study 2). The same dynamic was described by Kahan et al. (2012) in a study questioning the association commonly found between climate change beliefs and scientific literacy. That study shows that education, which can be assumed to be related to scientific literacy, is positively rather than negatively related to the degree of polarization concerning climate change beliefs between people holding different worldviews, which Kahan et al. (2012) interpreted to be a result of identity concerns. In this regard, Kahan et al. (2012) propose that “what guides individual risk
En savoir plus

33 En savoir plus

A Syntactic Possibilistic Belief Change Operator: Theory and empirical study

A Syntactic Possibilistic Belief Change Operator: Theory and empirical study

ical methods into a BDI architecture. Casali and col- leagues [7] proposed a general model for graded BDI agents and an architecture for modeling the agent’s graded mental attitudes. Like Casali, Blee and col- leagues [6] introduce levels in all the mentalistic no- tions of BDI, as well as using numeric, possibilistic- type functions in its semantics. We have presented an integrated theoretical framework, grounded in possi- bility theory, to account for all the aspects involved in representing and changing beliefs for cognitive agents [8]. In that framework, graded beliefs are rep- resented by means of a possibility distribution over in- terpretations, and a belief-change operator is proposed which obeys a possibilistic formulation of the AGM revision rationality postulates K∗1–K∗8 [16], and is a generalization of the possibilistic conditioning opera- tor of Dubois and colleagues [14]. Such operator is one of the members of a family of possibilistic condition- ing operators studied in [5].
En savoir plus

17 En savoir plus

Evaluating complex interventions: Perspectives and issues for health behaviour change interventions

Evaluating complex interventions: Perspectives and issues for health behaviour change interventions

(Barber, Connolly, Crits-Christoph, Gladis, & Siqueland, 2000 ) suggest that although early in treatment the alliance might be in fluenced by prior symptomatic improvement, it remains a signi ficant predictor of further improvement even when prior change in depression is partialled out. The role of the alliance as a potential causal factor in patient improvement is consistent with the theoretical and therapeutic role suggested by various psychodynamic and experiential theorists (Elvins & Green, 2008 ). Of all the variables studied, the therapeutic alliance has proved to be the best predictor (Krupnick et al., 1996 ). The correlation between level of alliance and symptomatic evolution of the patient is of the same order for all four groups (both psychotherapies, drug therapy and placebo group). Finally, the involvement of the alliance in outcome variance is much higher (about 21% of explained variance) than the therapeutic technique itself (only 2% of explained variance). But the alliance is not an intervention in and of itself.
En savoir plus

19 En savoir plus

Challenging Indifference to Extreme Poverty: Considering Southern Perspectives on Global Citizenship and Change

Challenging Indifference to Extreme Poverty: Considering Southern Perspectives on Global Citizenship and Change

Éthique et économique/Ethics and Economics, 8 (1), 2011, http://ethique-economique.net/ 118 would be the requirement of substantial preparation prior to going overseas, in the form of taking a pertinent course (or courses) for credit, for example, in the case of university and college students. The preparation aspect is very important, not only because of the potential to provide in-depth knowledge of the country the young person is going to, but also because this kind of awareness has the potential to affect motivations. If the postcolonial dimensions of the way the world operates are understood, and if there is awareness of the history and impact of national debt and structural adjustment programs, to name just two foundational areas for study, then it is more likely that a young person embarking on an international placement will do so with some sense of context for understanding the injustice of the poverty about to be encountered, some appreciation, even if only at an intellectual level, that it was not always so. That awareness might produce a stronger urge to make a positive contribution. I have argued elsewhere that the de-briefing process is equally important, for it is on the return to Canada that the meanings of experiences overseas become fully consolidated (Heron, 2005). Therefore, it would seem that requiring additional course work of students on their return would also be beneficial, provided that the course(s) undertaken enable critical self-reflection.
En savoir plus

11 En savoir plus

Internal models and private multi-agent belief revision

Internal models and private multi-agent belief revision

In the case of private announcement, the other agents’ belief clearly do not change. For example, suppose you (Y ) believe p, and agent j believes p (and perhaps even that p is common belief of Y and j). When a third external agent privately tells you that ¬p then j still believes p and you still believe that j believes p (and that j believes that p is com- mon belief). This static aspect of private announcements is similar to the static aspect of AGM belief revision in a single-agent case: in both cases the world does not change but only agent Y ’s beliefs about the world change. So, it is reasonable to expect that the AGM framework can be ex- tended to private multi-agent belief revision. In this paper we propose a natural generalization. The central device will be internal models, of which AGM models are a particular case.
En savoir plus

8 En savoir plus

A Logic of Explicit and Implicit Distributed Belief

A Logic of Explicit and Implicit Distributed Belief

The logical framework we presented is static, as the agents’ belief bases do not change. Future work will be devoted to bringing into our framework different types of belief change operations including belief base expansion, contraction and revision. The extended frame- work will allow us to study the interplay between belief change and belief aggregation. For instance, following [28, 30], we plan to inves- tigate whether it should be possible for the members of a coalition to come to individually believe that a certain fact ϕ is true, through in- formation exchange, when they distributively believe that ϕ.
En savoir plus

9 En savoir plus

A Combined System for Update Logic and Belief Revision

A Combined System for Update Logic and Belief Revision

Proposition: If we define K ∗ φ as the belief set associated with κ ∗ (φ, α), the revision function * thus defined satisfies the 8 AGM postulates. ⋄ So we have set out update logic and belief revision theory as viewed by W. Spohn. Now we are going to propose a system combining these two theories and see what insights it provides us regarding information change. As in the BMS exposition, we split our account in three parts: 1. Static part 2. Dynamic part 3. Update mechanism (inspired from W. Spohn’s theory).

18 En savoir plus

Combination of supervised and unsupervised classifiers based on belief functions

Combination of supervised and unsupervised classifiers based on belief functions

1.2.1 Difficulties in labeling process Sufficient labeled data with high quality is one of the prerequisites in land cover classi- fication. Despite various proposed land cover scheme, none of the current schemes has been internationally accepted because the definition of land covers greatly changes due to different purposes. The major process to define land cover scheme includes two phases: an initial dichotomous phase, and a subsequent modular-hierarchical phase, shown in Fig- ure 1.1. In the dichotomous phase, major land cover types (e.g., ice/snow, water, bare) are defined. In modular-hierarchical phase, environmental and technical attributes are incorporated in the definition to refine the major land cover types defined in the dichoto- mous phase. Environmental attributes (e.g., climate, altitude, time, and erosion) easily influence land covers possibly leading to mislabeled. Technical attributes (e.g., floristic aspect, crop types, and soils types) are also sometimes considered in defining land cover types for specific purposes. However, specific technical disciplines are often required dur- ing a ground survey, making the labeling process expensive and time-consuming. We take a recent land cover scheme proposed by NASA [142] as an example in Figure 1.2. The major land cover classes contain seven types: ice/snow, water, bare, developed, forest, herbaceous, and shrub. Below each of them, more refined class types are defined based on environmental attributes, marked in blue text, and/or specific technical attributes, marked in green text. Some classes related to environmental attributes, such as open shrub and sparse shrub, possibly change in a short time, and their boundaries become less clear to observe on the ground. Furthermore, specific knowledge is required to dis- tinguish phenological sub-classes of forest or shrub, types of bare soil, and rocks.
En savoir plus

191 En savoir plus

Belief Revision and the EM Algorithm

Belief Revision and the EM Algorithm

IRIT, CNRS and University of Toulouse, Toulouse, France Abstract. This paper provides a natural interpretation of the EM algo- rithm as a succession of revision steps that try to find a probability dis- tribution in a parametric family of models in agreement with frequentist observations over a partition of a domain. Each step of the algorithm cor- responds to a revision operation that respects a form of minimal change. In particular, the so-called expectation step actually applies Jeffrey’s re- vision rule to the current best parametric model so as to respect the frequencies in the available data. We also indicate that in the presence of incomplete data, one must be careful in the definition of the likeli- hood function in the maximization step, which may differ according to whether one is interested by the precise modeling of the underlying ran- dom phenomenon together with the imperfect observation process, or by the modeling of the underlying random phenomenon alone, despite imprecision.
En savoir plus

13 En savoir plus

Belief Update Revisited

Belief Update Revisited

1 Introduction In papers about belief change, one often reads sentences such as “belief revision consists in incorporating some new infor- mation about a static world, while belief update consists in in- corporating into a belief base about an old state of the world a notification of some change in the world”. However, the distinction is not so simple. To assess the scope of belief change operators, we need to be able to talk about the proper- ties of the system (the world and the available actions) and the properties of the agent’s state of knowledge, as in the taxon- omy for reasoning about action and change from [Sandewall, 1995]. However, unlike reasoning about action, belief change processes have never (as far as we know) been analyzed from the point of view of such a taxonomy. A first step is taken towards this direction (for belief revision only) in [Friedman and Halpern, 1996].
En savoir plus

6 En savoir plus

Belief Revision and the EM Algorithm

Belief Revision and the EM Algorithm

2 IRIT, CNRS and University of Toulouse, Toulouse, France Abstract. This paper provides a natural interpretation of the EM algo- rithm as a succession of revision steps that try to find a probability dis- tribution in a parametric family of models in agreement with frequentist observations over a partition of a domain. Each step of the algorithm cor- responds to a revision operation that respects a form of minimal change. In particular, the so-called expectation step actually applies Jeffrey’s re- vision rule to the current best parametric model so as to respect the frequencies in the available data. We also indicate that in the presence of incomplete data, one must be careful in the definition of the likeli- hood function in the maximization step, which may differ according to whether one is interested by the precise modeling of the underlying ran- dom phenomenon together with the imperfect observation process, or by the modeling of the underlying random phenomenon alone, despite imprecision.
En savoir plus

14 En savoir plus

Trust-based belief change

Trust-based belief change

On the technical level, the logic DL-BT consists of extending Liau’s static modal logic of belief and trust [12] in three different directions: (i) a generalization of Liau’s approach to graded trust, (ii) its extension by modal operators of knowledge and by modal oper- ators of graded belief based on Spohn’s theory of uncertainty [14], and (iii) by a family of dynamic operators in the style of dynamic epistemic logics (DEL) [16]. The latter allows for the representation of the consequences of a trust-based belief change operation while the second enables to handle iterated belief change.
En savoir plus

7 En savoir plus

Preference Change Triggered by Belief Change: a Principled Approach

Preference Change Triggered by Belief Change: a Principled Approach

1.5 Using our postulates To illustrate our postulates and their use to evaluate or classify preference change methods, we propose a general family of operators for preferences evolving after some new fact has been learned, parameterized by a revision function on epis- temic states and a semantics for interpreting preferences over formulas, and we give conditions on the revision function and the semantics of preference for each of our postulates to hold. We give here an informal presentation of the operators (formal details will be given in Section 2). In order to express preference revi- sion trigerred by belief revision we need to consider both relative plausibility be- tween worlds (or normality) and preference between worlds in our model. While a classical decision-theoretic approach would model plausibility and preference by probability distributions and utility functions respectively, we stick here to a purely ordinal modelling, following a long-standing tradition in the belief change community. Our models consist of two orderings on a set of worlds, one for nor- mality and one for preference, as illustrated on Figure 1. On the left hand side, the mental state of the agent is represented by two complete weak orders expressing respectively normality and preference, and new incoming information f results in the shift of f worlds towards normality, leaving the preference order unchanged. On the right hand side, the two complete weak orders are visualized more com- pactly by a two-dimensional structure. The striking out parts of the right hand side show the normality shift of the ¯ f -worlds.
En savoir plus

29 En savoir plus

Belief Change Based on Global Minimisation

Belief Change Based on Global Minimisation

time. (Note that, for each of these areas, there is obviously not a single nor a best way of minimising change.) Two questions come to mind. First, can we see (some of) these change-minimising approaches as instances of a more general framework? Second, are there other specific cases of this more general setting that are worth considering? This paper answers both questions. The general setting we con- sider starts with a set of points structured in a graph, and with a formula attached to each point. Following a minimisation step, a formula is determined at each point, representing the original information associated with the point along with in- formation gleaned from connected points. This minimisation step is driven by disagreements among the formulas attached to connected points. We show that a number of approaches to belief merging fit within this general setting, as do a number of approaches to reasoning about time and change, as well as an approach to revision.
En savoir plus

6 En savoir plus

Games with Communication: from Belief to Preference Change

Games with Communication: from Belief to Preference Change

The main originality of our contribution lies in the solution we offer for Question 2, by combining techniques and methods from game theory and from social choice theory [2]: informally, each possible strategy of B is seen as a voter, who votes for strategies of A according to the payoff A would obtain in the play defined by both strategies. Individual votes are then aggregated to define the new preferred strategy of A. Here again we do not choose a particular type of ballot nor a precise aggregation method, but rather leave it open and free to be set according to the kind of strategy one wants to obtain: for instance, one that has best average payoff against B’s most plausible strategies, or one that is most often a best response.
En savoir plus

9 En savoir plus

Show all 10000 documents...