AGISystem2 Research

Cognitive Architectures

Integration of symbolic and sub-symbolic modules into unified intelligence models.

Theoretical Foundations of Cognitive Architectures

Cognitive architectures provide frameworks for modeling human-like intelligence by integrating perception, memory, and reasoning into a unified system. These systems contrast with narrow AI by focusing on general-purpose problem solving.

Core Frameworks

Legacy and Specialized Architectures

Hybridization with Generative AI

Current research explores using LLMs as high-capacity associative memory modules within established cognitive architectures. This approach seeks to utilize LLMs for linguistic tasks while SOAR or ACT-R modules maintain long-term planning, logical consistency, and goal state monitoring.

References