Abstract and Keywords
ACT-R is a computational cognitive architecture intended to represent and simulate human thoughts and behaviors. It posits a declarative knowledge base with an associated computational account of memory recall and decay, and a procedural knowledge base with condition-action production rules representing procedural skill in performing task actions. By developing computational models using ACT-R, researchers can more rigorously explore the workings of the mind by predicting cognitive, perceptual, and motor behavior, and by comparing this behavior quantitatively to human behavior in the same tasks. ACT-R can also be used to generate quantitative predictions of behavior (e.g., on a human-computer interface) and infer a person’s underlying cognitive state for the duration of a task. ACT-R has played a central role in many research and applied efforts and now serves as a central repository of cognitive theory to which a community of cognitive researchers can contribute and from which they can benefit.