DataTerminology.comNEW
Also known as: Intelligent system architecture
A framework specifying the fixed structures and processes that generate intelligent behavior in an agent.
DESCRIPTION
Cognitive architectures define the building blocks of cognition—memory, perception, learning, and control—and how they interact. Examples include symbolic (Soar, ACT-R), connectionist, and hybrid approaches. They support modeling human cognition and building general intelligent agents. Architectural commitments guide representations, learning mechanisms, and reasoning strategies. Research compares architectures on explanatory breadth, computational performance, and psychological plausibility.
ALTERNATIVE DEFINITIONS
Wikipedia
A cognitive architecture is a blueprint for an intelligent agent, and a hypothesis about the fixed structures that provide a mind and yield intelligent behavior.
Positions architectures as both blueprints and scientific hypotheses.
Soar Project
Soar is a general cognitive architecture for developing systems that exhibit intelligent behavior.
Describes an instance of a cognitive architecture focused on general intelligent behavior.
RELATED TERMS
ACT-R Security Orchestration, Automation, and Response CLARION Production System Working Memory Long-Term Memory Cognitive Model Symbolic AI Neuro-Symbolic AI Reinforcement Learning
COMPETING TERMS
Connectionist Approach
RELATIONS
- Broader: Agent Architecture
KNOWLEDGE AREAS
TARGET AUDIENCE
AI Engineer AI Product Manager Enterprise Architect Data Architect
EXTERNAL SOURCES
- Wikipedia - Cognitive architecture
- ACT-R - ACT-R cognitive architecture site
- Soar - Soar architecture resources
- ScienceDirect - Topic overview and references
TOP QUESTIONS
Why do Cognitive Architectures matter for AI design?
They define the stable components and processes that constrain and enable intelligent behavior, guiding representation and learning choices.
Category: Relevancy
What distinguishes symbolic from connectionist Cognitive Architectures?
Symbolic architectures use explicit rules and symbols; connectionist approaches rely on distributed representations and learning in networks.
Category: Differences
How are Cognitive Architectures evaluated?
By explanatory scope, psychological plausibility, computational performance, and ability to generalize across tasks.
Category: Testing
Where should I start learning Cognitive Architectures?
Explore ACT-R and Soar sites for tutorials and models, then survey hybrid and neurosymbolic approaches.
Category: Getting Started
What risks exist when choosing a Cognitive Architecture?
Overcommitting to an architecture can limit flexibility; mismatched assumptions may impair learning or scalability.
Category: Risks
How do Cognitive Architectures support explainability?
They often encode interpretable components (e.g., productions, memories), aiding traceable reasoning and explanations.
Category: Application
What is a hot topic in Cognitive Architectures?
Hybrid and neurosymbolic systems integrating reasoning with deep learning are gaining traction.
Category: Topic
Can Cognitive Architectures model human behavior?
Yes, architectures like ACT-R model cognitive tasks and predict performance, informing HCI and training design.
Category: Example
How do Cognitive Architectures relate to AGI efforts?
They provide candidate blueprints for general intelligence, though debates persist on required modules and learning capacities.
Category: Opinion