Connectionist "face"-off: Different algorithms for different tasks

Abstract

We present a series of simulations contrasting the ability of a Hebbian and a Widrow-Hoff trained autoassociative memory to perform several face processing tasks in different learning and testing conditions. We show that all face processing tasks are not equally demanding, and that, in some particular circumstances, a simple learning algorithm can be more appropriate than u complex one. We also illustrate that the choice of evaluation criteria for assessing the performance of a model is crucially dependent on the task performed, especially when trying to predict human behavior. Finally, we speculate about the complexity of early learning (i.e., by infants) of faces, and suggest that this task is easier than has been generally thought.