v1.0: Add Zenodo DOI, cleaned paper PDF

.gitattributes

@@ -48,3 +48,4 @@ static/videos/teaser.mp4 filter=lfs diff=lfs merge=lfs -text
 static/videos/toby.mp4 filter=lfs diff=lfs merge=lfs -text
 papers/fig1_optimization_trajectory.png filter=lfs diff=lfs merge=lfs -text
 papers/fig2_snr_analysis.png filter=lfs diff=lfs merge=lfs -text
+papers/The[[:space:]]Atomic[[:space:]]VSA.pdf filter=lfs diff=lfs merge=lfs -text

README.md

@@ -8,21 +8,24 @@ pinned: false
 license: mit
 ---
 
-# Atomic VSA: A Breakthrough in Deterministic, High-Fidelity AI
+# Atomic VSA: Physics-Inspired Hyperdimensional Computing for Explainable AI
+
+[![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.18650281.svg)](https://doi.org/10.5281/zenodo.18650281)
+[![GitHub](https://img.shields.io/badge/GitHub-atomic--vsa--research-blue)](https://github.com/muhammadarshad/atomic-vsa-research)
 
 **Author:** Muhammad Arshad  
 **Date:** February 15, 2026
 
-This repository contains the official implementation and reproduction scripts for the paper **"The Atomic VSA: A Breakthrough in Deterministic, High-Fidelity AI"**.
-
-The Atomic Vector Symbolic Architecture (Atomic VSA) is a deterministic AI framework that unifies Holographic Algebra with Inverse Frequency physics to resolve the **Accuracy vs. Efficiency vs. Interpretability** trilemma in clinical triage.
+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
 
-- **98.4% F1 Score** on clinical datasets.
-- **Microsecond Latency** (O(1)) on standard CPUs.
-- **Full Interpretability** using transparent algebraic operations.
-- **Green AI**: Eliminates the need for massive GPU clusters.
+| 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
 

papers/The Atomic VSA.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cb39eba51b0d553ec2215dc53072ad2d62d2365775ba009f367cd4d948cdb226
+size 1144587

papers/cite.cff

@@ -3,32 +3,34 @@ message: "If you use this software or methodology, please cite it as below."
 authors:
   - family-names: "Arshad"
     given-names: "Muhammad"
+    email: "marshad.dev@gmail.com"
     country: "PK"
-    email: "marshad.dev@gmail.com" 
-title: "The Atomic VSA: A Breakthrough in Deterministic, High-Fidelity AI"
+title: "Atomic Vector Symbolic Architecture: Physics-Inspired Hyperdimensional Computing for Explainable AI"
 version: 1.0.0
 date-released: 2026-02-15
-url: "https://huggingface.co/spaces/marshad180/Atomic-VSA"
+doi: "10.5281/zenodo.18650281"
+repository-code: "https://github.com/muhammadarshad/atomic-vsa-research"
+url: "https://zenodo.org/records/18650281"
 preferred-citation:
   type: article
   authors:
     - family-names: "Arshad"
       given-names: "Muhammad"
-      country: "PK"
-  title: "The Atomic VSA: A Breakthrough in Deterministic, High-Fidelity AI"
+  title: "Atomic Vector Symbolic Architecture: Physics-Inspired Hyperdimensional Computing for Explainable AI"
   year: 2026
   month: 2
   status: preprint
-  url: "https://huggingface.co/spaces/marshad180/Atomic-VSA"
 keywords:
   - "Vector Symbolic Architectures"
-  - "Deterministic AI"
+  - "Hyperdimensional Computing"
   - "Clinical Decision Support"
+  - "Explainable AI"
+  - "Deterministic AI"
   - "Green AI"
-  - "Hyperdimensional Computing"
 abstract: >
   The Atomic Vector Symbolic Architecture (Atomic VSA) is a deterministic AI framework 
-  that unifies Holographic Algebra with Inverse Frequency physics to resolve the 
-  Accuracy vs. Efficiency trilemma in clinical triage. It achieves 98.4% F1 scores 
-  with O(1) inference latency on standard CPUs, eliminating the need for massive 
-  GPU clusters.
+  grounded in hyperdimensional computing. Using 10,048-dimensional bipolar vectors with 
+  thermometer encoding and IDF weighting, the system achieves 92.5% F1 on 25-category 
+  clinical triage with 11.97ms inference latency on commodity CPUs. The framework 
+  provides algebraically invertible operations enabling full explainability without 
+  post-hoc interpretation methods.