Decision theory and artificial intelligence II: The hungry monkey

This paper describes a problem-solving framework In which aspects of mathematical decision theory are incorporated into symbolic problem-solving techniques currently predominant in artificial intelligence. The utility function of decision theory IS used to reveal tradeoffs among competing strategies for achieving various goals, taking into account such factors as reliability, the complexity of steps in the strategy, and the value of the goal. The utility function on strategies can therefore be used as a guide when searching for good strategies. It is also used to formulate solutions to the problems of how to acquire a world model, how much planning effort is worthwhile, and whether verification tests should be performed. These techniques are illustrated by application to the classic monkey and bananas problem.     

Mapping the mental space of rectangles.

The cognitive structure of a shape space--the space of rectangles--is explored by a nonmetric scaling technique. Our experiment was designed to extract the major transformational paths or 'modes' that characterize the mental shape space. Earlier studies of rectangle similarities using multidimensional scaling have provided conflicting evidence about whether the coordinate system of the mental rectangle space is based on height and width or on area and shape (i.e. aspect ratio). Our study reveals shape to be the single dominant factor. We suspected that earlier evidence for a height-width parameterization might have been due to the presentation of rectangles upright in a pseudo-gravitational coordinate system (whereas our rectangles are randomly rotated). In a control experiment with upright (vertical or horizontal) rectangles, the heavy bias towards shape preservation was still the dominant mode. In addition, however, a secondary bias towards change of height or width emerged, exactly following the pattern expected from the biasing change in context. This finding established a concrete path by which context and frame can influence the way shape is represented. The relevance of these findings to the cognitive organization of more complex shape spaces is discussed.     

Phonological ambiguity and lexical ambiguity: effects on visual and auditory word recognition

Three experiments in Serbo-Croatian were conducted on the effects of phonological ambiguity and lexical ambiguity on printed word recognition. Subjects decided rapidly if a printed and a spoken word matched or not. Printed words were either phonologically ambiguous (two possible pronunciations) or unambiguous. If phonologically ambiguous, either both pronunciations were real words or only one was, the other being a nonword. Spoken words were necessarily unambiguous. Half the spoken words were auditorily degraded. In addition, the relative onsets of speech and print were varied. Speed of matching print to speech was slowed by phonological ambiguity, and the effect was amplified when the stimulus was also lexically ambiguous. Auditory degradation did not interact with print ambiguity, suggesting that perception of the spoken word was independent of the printed word.     

Psychosocial and pharmacological treatment of patients following deliberate self-harm: the methodological issues involved in evaluating effectiveness.

Development of effective treatments for patients following deliberate self-harm (self-poisoning or self-injury) is a very important element in suicide prevention. The randomized controlled trial (RCT) is the mainstay of evaluation of treatments. In a systematic review of the literature, the effectiveness of treatments based on RCTs was examined and the quality of the RCTs was assessed. Twenty trials were identified, and where possible, these were grouped on the basis of similarities among the types of treatment. In this paper, we examine the methodological aspects of the trials and consider what may be learned that will assist in the design of future studies in this field. The methodological quality of the trials was reasonable, but most trials included too few participants to detect clinically important differences in rates of repeated self-harm. In planning future trials, the following major issues should be addressed: investigators should perform power calculations to determine the number of subjects necessary to detect clinically important effects, provide information on method of randomization and interventions, use standard measures of outcome, and focus on homogeneous subgroups of patients. Improving the methodology of future studies in this field will be essential if sound evidence is to be obtained which can inform effective service provision for deliberate self-harm patients.