ACT-R

A modular cognitive architecture from Carnegie Mellon University designed to simulate human thought processes through rule-based symbolic and subsymbolic systems.

John Anderson's team at Carnegie Mellon University developed this architecture to model the discrete steps of human cognition. The system functions through a series of modules: declarative memory stores facts (chunks) while procedural memory executes rules (productions). It operates on a 50-millisecond cycle to mirror actual human neural processing speeds. Developers use ACT-R to build intelligent tutors for mathematics and high-fidelity pilot simulations for the Department of Defense. It translates psychological theory into executable code to predict how humans learn, forget, and react under pressure.

https://act-r.psy.cmu.edu/