README.md

metadata

title: Atomic VSA
emoji: ⚛️
colorFrom: blue
colorTo: indigo
sdk: docker
pinned: false
license: mit

Atomic VSA: Physics-Inspired Hyperdimensional Computing for Explainable AI

Author: Muhammad Arshad
Date: February 15, 2026

This Space hosts the official implementation of the Atomic Vector Symbolic Architecture (Atomic VSA) — a deterministic framework grounded in Hyperdimensional Computing (HDC) for clinical triage and explainable AI.

🚀 Key Results

Metric	Result
F1 Score	92.5% (25-category ICD-11 triage)
Label Recall	91.9% (multi-label comorbidity)
Inference Latency	11.97ms (p50, commodity CPU)
Power Consumption	15W (edge deployment)

📂 Repository Structure

• src/: Core implementation in Julia.
• scripts/: Python scripts for reproducing paper figures.
• papers/: The full research paper (PAPER_7_ATOMIC_VSA_BREAKTHROUGH.md), LaTeX source, and generated figures.

🛠️ Usage

Python (Reproduction Scripts)

1. Install dependencies:

   pip install -r requirements.txt
   

2. Generate paper charts:

   python scripts/generate_paper_charts.py
   

   The charts will be saved to the papers/ directory.

Julia (Core Logic)

The core logic is implemented in Julia. You can explore the src/ directory to see the implementation of the Atomic algebra, VortexEngine, and other components.

📜 Citation

If you use this work, please cite it:

@article{arshad2026atomicvsa,
  title={The Atomic VSA: A Breakthrough in Deterministic, High-Fidelity AI},
  author={Arshad, Muhammad},
  year={2026},
  publisher={Hugging Face}
}

See papers/cite.cff for more citation formats.