CEREBRUM

CEREBRUM: Case-Enabled Reasoning Engine with Bayesian Representations for Unified Modeling

Repository

GitHub Repository: https://github.com/ActiveInferenceInstitute/CEREBRUM

Publication Information

• Author: Daniel Ari Friedman
• Version: 1.4
• Date: April 7, 2025
• DOI: 10.5281/zenodo.15170908
• Zenodo Record: https://zenodo.org/records/15231156
• Institution: Active Inference Institute

Overview

CEREBRUM is a unified modeling framework that integrates case-based reasoning with Bayesian representations. It provides a comprehensive suite of tools for developing, testing, and deploying cognitive models that combine the strengths of case-based reasoning with probabilistic inference.

Key features include:

• Integration of symbolic case structures with Bayesian probabilistic representations
• Flexible transformation engine for model adaptation and learning
• Multiple language implementations with consistent APIs
• Visualization tools for model inspection and analysis
• Extensive example library demonstrating applications across domains

Project Architecture

graph TD
    A[CEREBRUM Framework] --> B[Core Engine]
    A --> C[Model Library]
    A --> D[Transformation Engine]
    A --> E[Visualization Tools]
    
    B --> B1[Active Inference Module]
    B --> B2[Bayesian Representation]
    B --> B3[Case Management]
    
    C --> C1[Predefined Models]
    C --> C2[Custom Model Templates]
    
    D --> D1[Model Adaptation]
    D --> D2[Case Transformation]
    D --> D3[Learning Mechanisms]
    
    E --> E1[Interactive Visualizations]
    E --> E2[Model Inspection Tools]
    E --> E3[Diagnostic Reports]
    
    style A fill:#f9f,stroke:#333,stroke-width:2px
    style B fill:#bbf,stroke:#333
    style C fill:#bbf,stroke:#333
    style D fill:#bbf,stroke:#333
    style E fill:#bbf,stroke:#333

Loading

Repository Structure

The CEREBRUM repository is organized into the following structure:

Directory/File	Description
/src	Source code for the CEREBRUM framework
/src/core	Core engine components including active inference module, neural networks, and model base classes
/src/models	Implementation of various model types and case definitions
/src/utils	Utility functions and helper classes
/src/visualization	Visualization tools and components
/src/transformations	Model transformation and adaptation algorithms
/src/cases	Case definition and management functionalities
/src/examples	Example implementations and demonstrations
/docs	Documentation including specifications, guides, and discussions
/docs/examples	Detailed documentation for example models
/docs/languages	Language-specific implementation guidelines
/paper	Academic paper source files and generation tools
/paper/components	Paper section components and figures
/paper/output	Generated paper outputs in various formats
/tools	Development and maintenance tools
/output	Default location for generated outputs and results

Implementation Workflow

flowchart LR
    A[Define Case Structure] --> B[Create Bayesian Model]
    B --> C[Configure Transformations]
    C --> D[Train/Adapt Model]
    D --> E[Evaluate Results]
    E -->|Iterate| A
    
    style A fill:#bbf,stroke:#333
    style B fill:#bbf,stroke:#333
    style C fill:#bbf,stroke:#333
    style D fill:#bbf,stroke:#333
    style E fill:#bbf,stroke:#333

Loading

Getting Started

CEREBRUM can be used in multiple ways depending on your needs:

1. Explore existing models: Check out the /src/examples directory
2. Create your own models: Follow the guides in /docs/getting_started.md
3. Extend the framework: See /docs/contributing_technical.md for development setup

Installation

# Clone the repository
git clone https://github.com/ActiveInferenceInstitute/CEREBRUM.git
cd CEREBRUM

# Install dependencies
pip install -r requirements.txt

# For development
pip install -r requirements_dev.txt

PDF Documentation

The full documentation for CEREBRUM is available as a PDF file in the paper/output/ directory, which is generated from the Markdown source files.

Requirements for PDF Generation

To regenerate the PDF documentation, you need:

• Python 3
• Node.js with Mermaid CLI (@mermaid-js/mermaid-cli package)
• Pandoc
• XeLaTeX (for PDF generation)

Generating the PDF

To regenerate the PDF documentation, run:

python3 paper/assemble_paper.py

This script will:

1. Render all Mermaid diagrams from paper/components/figures/ to PNG images
2. Assemble all markdown components in the correct order
3. Generate an HTML version of the paper
4. Create the final PDF file at paper/output/assembled_paper.pdf

For more detailed information about the paper building process, please see paper/README.md.

Use Cases

CEREBRUM has been applied to a variety of domains:

mindmap
  root((CEREBRUM))
    Cognitive Science
      Mental Models
      Decision Making
      Learning Processes
    Artificial Intelligence
      Agent Modeling
      Reinforcement Learning
      Multi-agent Systems
    Control Systems
      HVAC Controls
      Robotics
      Adaptive Systems
    Natural Language
      Text Understanding
      Semantic Processing
      Dialog Systems

Loading

Citation

If you use CEREBRUM in your research, please cite:

@misc{friedman2025cerebrum,
  author = {Friedman, Daniel Ari},
  title = {CEREBRUM: Case-Enabled Reasoning Engine with Bayesian Representations for Unified Modeling},
  year = {2025},
  month = {April},
  version = {1.4},
  doi = {10.5281/zenodo.15170908},
  publisher = {Active Inference Institute},
  url = {https://github.com/ActiveInferenceInstitute/CEREBRUM}
}

License

CC BY-NC-ND 4.0

Acknowledgements

This work builds upon the efforts and insights of many participants within the Active Inference Institute. Special thanks are extended to Dave Douglass for his contributions related to computational linguistics, archiving, Active Inference principles, upper ontologies, and translations, which have significantly informed aspects of this project.

Contributing

We welcome contributions from researchers, developers, and enthusiasts across multiple domains:

Ways to Contribute

• 🔬 Research: Theoretical development, empirical testing, literature reviews
• 💻 Technical: Core development, language implementations, integrations, testing
• 📚 Documentation: Educational content, API docs, use cases, tutorials
• 🧩 Examples: Model examples, domain applications, interactive demonstrations
• 🌍 Community: Organizing events, content creation, mentoring, user support

Contribution Scale

From micro-contributions (1-3 hours) to large projects (40+ hours), we value contributions of all sizes.

Development Process

graph LR
    A[Fork Repository] --> B[Setup Dev Environment]
    B --> C[Create Feature Branch]
    C --> D[Implement Changes]
    D --> E[Run Tests]
    E -->|Pass| F[Submit PR]
    E -->|Fail| D
    F --> G[Code Review]
    G -->|Approved| H[Merge]
    G -->|Changes Requested| D
    
    style A fill:#bbf,stroke:#333
    style H fill:#9f9,stroke:#333

Loading

For detailed contribution guidelines, process information, and more specific ideas, please see our CONTRIBUTING.md document.