Analogical Inferences in the Family Trees Task: A Review
We reviewed the performance results on a generalization test obtained by Hinton in a pioneering study of perceptrons’ capacity to make analogical inferences. In that test, a five-layer feed-forward network was exposed to items that were not part of its training set (i.e., four relationships between members of two genealogically identical fa-milies), to see whether it could nevertheless generalize and produce the correct output. The good performance showed by two simulations was interpreted by Hinton as an evidence of its ability to detect the common structure shared by both families and make analogical inferences from one family tree to the other. However, we claim that this work’s good results were clouded by unsuitable test items, and that it also lacked a control condition to prove that analogical inferences were actually done. Our study, which tackled these two aspects through experimental manipulation, provides stronger evidence supporting Hinton’s conclusions.
KeywordsAnalogical inference structure sensitivity systematicity
Unable to display preview. Download preview PDF.
- 1.Hinton, G.E.: Learning Distributed Representations of Concepts. In: Proceedings of the Eight Annual Conference of the Cognitive Science Society, pp. 1–12. Lawrence Erlbaum Associates, New Jersey (1986)Google Scholar
- 3.McLeod, P., Plunkett, K., Rolls, E.T.: Introduction to Connectionist Modelling of Cognitive Processes. Oxford University Press, New York (1998)Google Scholar
- 4.Melz, E.R.: Developing Microfeatures by Analogy. In: Proceedings of the Fourteenth Annual Conference of the Cognitive Science Society, pp. 42–47. Lawrence Erlbaum Associates, New Jersey (1992)Google Scholar