Enhanced ground support: lessons from work on reduced crew operations

Cognition, Technology & Work - From the 1950s through the 1980s, aircraft design was marked by an increase in reliability and automation, and, correspondingly, a decrease in the crew complement...     

A production system theory of serial memory.

A theory is described that provides a detailed model of how people recall serial lists of items. This theory is based on the Adaptive Character of Thought-Rational (ACT-R) production system (J. R. Anderson, 1993). It assumes that serial lists are represented as hierarchical structures consisting of groups and items within groups. Declarative knowledge units encode the position of items and of groups within larger groups. Production rules use this positional information to organize the serial recall of a list of items. In ACT-R, memory access depends on a limited-capacity activation process, and errors can occur in the contents of recall because of a partial matching process. These limitations conspire in a number of ways to produce the limitations in immediate memory span. As the span increases, activation must be divided among more elements, activation decays more with longer recall times, and there are more opportunities for positional and acoustic confusions. The theory is shown to be capable of predicting both latency and error patterns in serial recall. It addresses effects of serial position, list length, delay, word length, positional confusion, acoustic confusion, and articulatory suppression. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Explorations of an Incremental,Bayesian Algorithm for Categorization

An incremental categorization algorithm is described which, at each step, assigns the next instance to the most probable category. Probabilities are estimated by a Bayesian inference scheme which assumes that instances are partitioned into categories and that within categories features are displayed independently and probabilistically. This algorithm can be shown to be an optimization of an ideal Bayesian algorithm in which predictive accuracy is traded for computational efficiency. The algorithm can deliver predictions about any dimension of a category and does not treat specially the prediction of category labels. The algorithm has successfully modeled much of the empirical literature on human categorization. This paper describes its application to a number of data sets from the machine learning literature. The algorithm performs reasonably well, having its only serious difficulty because the assumption of independent features is not always satisfied. Bayesian extensions to deal with nonindependent features are described and evaluated.