Holographic Declarative Memory and the Fan Effect: A Test Case for A New Memory Module for ACT-R

We present Holographic Declarative Memory (HDM), a newmemory module for ACT-R and alternative to ACT-R’sDeclarative Memory (DM). ACT-R is a widely used cognitivearchitecture that models many different aspects of cognition,but is limited by its use of words as symbols to representideas or stimuli. HDM replaces the symbols with holographicvectors. Holographic vectors retain the expressive power ofsymbols but have a similarity metric, allowing for shades ofmeaning, fault tolerance, and lossy compression. The purposeof HDM is to enhance ACT-R’s ability to learn associations,learn over the long-term and store large quantities of data, anduse partial or fuzzy matching. To demonstrate HDM, we fitperformance of an ACT-R model that uses HDM to abenchmark memory task, the fan effect.