Academic papers and technical whitepapers for the Are-Self project — an open-source, neuro-mimetic AI reasoning engine that runs on consumer hardware.
| Paper | Status | Authors | Description |
|---|---|---|---|
| Neuro-Mimetic Architecture | Draft | Michael Clark | The flagship paper. Describes the brain-region software architecture, tick-cycle execution model, addon pipeline, and real-time event system. |
| The Focus Economy | Draft | Michael Clark | Bounded autonomy through incentive-aligned resource management. How linking execution budgets to memory novelty prevents degenerate agent behavior. |
| Hippocampus Hypergraph Migration | Draft | Samuel Frerichs, Michael Clark | Proposed migration from flat Engram M2M relations to a hypergraph memory model with typed relationships (causal, temporal, contradicts, etc.). |
| LLM Testing Harness | Draft | Michael Clark | Evaluating local models for agent-specific criteria: tool call compliance, multi-turn consistency, focus efficiency, memory formation quality, and quantization effects. |
| CI/CD Sovereignty | Draft | Michael Clark | Using Are-Self's neural pathway system as a self-hosted CI/CD orchestrator with AI-assisted failure diagnosis. |
| Multiplayer Unreal + Autonomous AI | Draft | Michael Clark | Case study of building a multiplayer Unreal Engine project with Are-Self as build engineer and project manager. |
These papers describe the architecture, theory, and evaluation of the Are-Self system. The code lives in:
- are-self-api — Django backend (brain regions as Django apps)
- are-self-ui — React frontend (graph editor, monitor, reasoning view)
- are-self-docs — Documentation site at are-self.com
Papers are written in LaTeX using the APA 7th edition format (apa7 document class). See templates/are-self-paper.tex for the base template.
cd papers/neuro-mimetic-architecture
pdflatex paper.tex
biber paper
pdflatex paper.tex
pdflatex paper.texOr use Overleaf for collaborative editing — import the paper directory and it compiles in the browser.
- A LaTeX distribution (TeX Live, MiKTeX, or MacTeX)
- The
apa7document class (tlmgr install apa7) - Biber for bibliography processing
These papers are living documents that evolve alongside the codebase. If you're interested in contributing research, see the Contributing Guide and open an issue describing your proposed contribution.
Papers are licensed under CC BY 4.0. Code examples within papers are MIT licensed, consistent with the main project.
