---
id: cognitive-architectures
title: "Cognitive Architectures: ACT-R, Soar, and Computational Models of Human-Like Reasoning"
schema_type: article
category: ai
language: en
confidence: medium
last_verified: "2026-05-28"
created_date: "2026-05-24"
generation_method: ai_structured
ai_models:
  - claude-4.5-sonnet
derived_from_human_seed: true
conflict_of_interest: none_declared
is_live_document: false
data_period: static
completeness: 0.85
atomic_facts:
  - id: af-cognitive-architectures-1
    statement: >-
      The ACT-R project develops a cognitive architecture and applies it to modeling human interaction with complex
      environments.
    source_title: ACT-R | Human-Computer Interaction Institute
    source_url: https://hcii.cmu.edu/project/act-r
    confidence: medium
  - id: af-cognitive-architectures-2
    statement: >-
      The Soar documentation distinguishes the architecture itself from the long-term knowledge added to build a
      particular Soar program.
    source_title: The Soar Architecture
    source_url: https://soar.eecs.umich.edu/soar_manual/02_TheSoarArchitecture/
    confidence: medium
  - id: af-cognitive-architectures-3
    statement: >-
      The Soar Cognitive Architecture gives a book-length account of Soar as both a software system and a theory of
      cognitive structure.
    source_title: The Soar Cognitive Architecture
    source_url: https://mitpress.mit.edu/9780262538534/the-soar-cognitive-architecture/
    confidence: medium
primary_sources:
  - title: ACT-R | Human-Computer Interaction Institute
    type: academic_project
    year: 2026
    institution: Carnegie Mellon University
    url: https://hcii.cmu.edu/project/act-r
  - title: The Soar Architecture
    type: technical_documentation
    year: 2026
    institution: University of Michigan
    url: https://soar.eecs.umich.edu/soar_manual/02_TheSoarArchitecture/
  - title: The Soar Cognitive Architecture
    type: book
    year: 2012
    institution: MIT Press
    url: https://mitpress.mit.edu/9780262538534/the-soar-cognitive-architecture/
known_gaps:
  - Scaling laboratory cognitive models to open-ended real-world tasks
  - Validation against neural, behavioral, and interaction data
disputed_statements: []
secondary_sources: []
updated: "2026-05-28"
---
## TL;DR
Cognitive architectures are computational frameworks for modeling integrated cognition: memory, goals, perception, action, learning, and problem solving. They are different from ordinary task models because they aim to explain how many cognitive functions work together.

## Core Explanation
ACT-R and Soar are two influential examples. ACT-R is used to model human cognition with interacting memory and production systems. Soar is a general architecture for building intelligent agents and studying the mechanisms needed for human-level behavior.

## Detailed Analysis
The research value is explanatory as much as performance-oriented. A cognitive architecture can generate predictions about behavior, response time, or errors, while also providing an implementation for synthetic agents. The hard part is scaling models without losing interpretability or empirical grounding.

## Further Reading
- ACT-R project
- Soar project
- The Soar Cognitive Architecture

## Related Articles

- [3D Human Modeling: Parametric Body Models, Mesh Recovery, and Digital Avatars](../3d-human-modeling.md)
- [Agentic AI: Autonomous Agent Architectures, Planning, and Tool-Integrated Reasoning](../agentic-ai.md)
- [Computational Neuroscience: AI Models of Brain Circuits, Connectomics, and Neural Computation](../computational-neuroscience.md)