A Hybrid Connectionist-Instance Model of Recognition Memory

While connectionist models have been proposed and met with some success in many areas of psychology, connectionist models of memory have not met the standards created by competing models of memory. Much of the reason for this failure is due to the severe interference connectionist models typically show for information learned sequentially, which is often called catastrophic interference. Solutions to the catastrophic interference problem in turn sacrifice the properties that make these models attractive. This dissertation proposes an alternative connectionist model of recognition memory, which contains a solution to the catastrophic interference problem that preserves the strengths of the connectionist approach. This model not only solves the catastrophic interference problem, but proposes a unique solution to the need for models of recognition to explain both interference and generalization processes in memory.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Psychology

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2015

Author dc:creator

• Adams, David Russell

Contributors dc:contributor

• Dell, Gary S.

• Psychology, Cognitive

Language dc:language

eng

Identifier

(MiAaPQ)AAI9812517

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/82217