Concurrent Knowledge Activation Calculation in Large Declarative Memories

To behave effectively and flexibly in complex situations, models specified in cognitive architectures must be able to store and access large amounts of declarative knowledge. However, as research efforts employing cognitive modeling grow in scope and complexity, currently available modeling tools, languages and cognitive architectures are being pushed to their practical limits. This paper describes research looking specifically at how a large declarative memories challenge the current implementation of ACT-R and describes an applied effort to develop an alternative implementation of ACT-R’s retrieval process. The alternative exploits concurrency features of the Erlang programming language to extend the practicality of ACT-R’s retrieval mechanisms to new levels of scale. The ideas and methods underlying the alternative implementation are general and illustrate how concurrency can accelerate calculation in other architectures struggling to support large associative declarative memories.