Computing explanations for interactive constraint-based systems

Constraint programming has emerged as a successful paradigm for modelling combinatorial problems arising from practical situations. In many of those situations, we are not provided with an immutable set of constraints. Instead, a user will modify his requirements, in an interactive fashion, until he is satisfied with a solution. Examples of such applications include, amongst others, model-based diagnosis, expert systems, product configurators. The system he interacts with must be able to assist him by showing the consequences of his requirements. Explanations are the ideal tool for providing this assistance. However, existing notions of explanations fail to provide sufficient information. We define new forms of explanations that aim to be more informative. Even if explanation generation is a very hard task, in the applications we consider, we must manage to provide a satisfactory level of interactivity and, therefore, we cannot afford long computational times. We introduce the concept of representative sets of relaxations, a compact set of relaxations that shows the user at least one way to satisfy each of his requirements and at least one way to relax them, and present an algorithm that efficiently computes such sets. We introduce the concept of most soluble relaxations, maximising the number of products they allow. We present algorithms to compute such relaxations in times compatible with interactivity, achieving this by indifferently making use of different types of compiled representations. We propose to generalise the concept of prime implicates to constraint problems with the concept of domain consequences, and suggest to generate them as a compilation strategy. This sets a new approach in compilation, and allows to address explanation-related queries in an efficient way. We define ordered automata to compactly represent large sets of domain consequences, in an orthogonal way from existing compilation techniques that represent large sets of solutions.

Deriving explanations from partial temporal information

The representation and manipulation of natural human understanding of temporal phenomena is a fundamental field of study in Computer Science, which aims both to emulate human thinking, and to use the methods of human intelligence to underpin engineering solutions. In particular, in the domain of Artificial Intelligence, temporal knowledge may be uncertain and incomplete due to the unavailability of complete and absolute temporal information. This paper introduces an inferential framework for deriving logical explanations from partial temporal information. Based on a graphical representation which allows expression of both absolute and relative temporal knowledge in incomplete forms, the system can deliver a verdict to the question if a given set of statements is temporally consistent or not, and provide understandable logical explanation of analysis by simplified contradiction and rule based reasoning.

We could make a meal out of nothing: contested explanations of inequalities in health

This paper reports on the qualitative findings from a comparative study of public health and lifestyles in South East England and Northern France, regions with similar geographic and economic characteristics. Data from health surveys showed that both countries had an increasing BMI with age, particularly in Northern France. This was despite the finding that the percentage eating fresh fruit and vegetable at least five days a week in Northern France increased with age (from well over 50% to over 90%) compared to around 50% to around 75% in South East England. Qualitative data on health inequalities and how they could be addressed were gathered by focus groups sampling from five tiers using the Townsend Index for comparability (14 in England with 106 participants overall; 13 in France with 143 participants). Both had about two thirds women participants, with a preponderance of middle aged and older people. There was a striking difference in the salience of diet between the two countries; in the French data it was raised only 14 times, whereas in England there were 165 occurrences, and these were often distinguished by their use of narrative. Older respondents contrasted the pressures on families today and the expense of fresh fruit and vegetables with their own childhood or childrearing, when cheap meals could be created using skills which have now been lost. These data therefore provide further evidence that providing food is a moral activity.

Explanations for child sexual abuse given by convicted offenders in Malawi:no evidence for "HIV cleansing"

OBJECTIVE: A commonly cited, but unproven reason given for the rise in reported cases of child sexual abuse in Sub-Saharan Africa is the "HIV cleansing myth"-the belief that an HIV infected individual can be cured by having sex with a child virgin. The purpose of this study was to explore in Malawi the reasons given by convicted sex offenders for child sexual abuse and to determine if a desire to cure HIV infection motivated their offence.

METHODS: Offenders convicted of sexual crimes against victims under the age of 18 were interviewed in confidence in Malawi's two largest prisons. During the interview the circumstances of the crime were explored and the offenders were asked what had influenced them to commit it. Each participant was asked the closed question "Did you think that having sex with your victim would cure or cleanse you from HIV?"

RESULTS: 58 offenders agreed to participate. The median (range) age of offenders and victims was 30 (16-66) years and 14 (2-17) years, respectively. Twenty one respondents (36.2%) denied that an offence had occurred. Twenty seven (46.6%) admitted that they were motivated by a desire to satisfy their sexual desires. Six (10.3%) stated they committed the crime only because they were under the influence of drugs or alcohol. None of the participants said that a desire to cure or avoid HIV infection motivated the abuse.

CONCLUSION: This study suggests that offenders convicted of a sexual crime against children in Malawi were not motivated by a desire to be cured or "cleansed" from HIV infection. A need to fulfil their sexual urges or the disinhibiting effect of drugs or alcohol was offered by the majority of participants as excuses for their behaviour.

The Influence of Causal Explanations and Diagnostic Labeling on Psychology Students’ Beliefs About Treatments, Prognosis, Dangerousness and Unpredictability in Schizophrenia

This study explored views of 566 Italian psychology students about schizophrenia. The most frequently cited causes were psychological traumas (68%) and heredity (54%). Thirty-three percent of students firmly believed that people with the condition could recover. Reporting heredity among the causes, and identifying schizophrenia were both associated with prognostic pessimism, greater confidence in pharmacological treatments and lower confidence in psychological treatments. Schizophrenia labeling was also associated with higher perception of unpredictability and dangerousness. Compared to first year students, fourth/fifth year students more frequently reported heredity among the causes, and were more pessimistic about schizophrenia recovery. Stigma topics should be included in future psychologists’ education.