Skip to content

avatariamfoundation-afk/neurogrid-core

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

395 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Quick Start for Judges

➡️ Start here:
SYSTEM_OVERVIEW.md

This one-page document explains the full NeuroGrid system, live deployment, and design intent in under 2 minutes.


CI Status (Hackathon Note)

Continuous Integration workflows are intentionally disabled during the hackathon phase. Reason:

  • Core logic is frozen for audit
  • MedIntel and DeSci are non-executing by design
  • CI re-enabled post-hackathon with full test coverage

This is a deliberate, documented decision.

Audit Status: Core logic frozen under tag v0.1.0-hackathon-freeze


NeuroGrid-Core (Summary)

NeuroGrid is a deterministic coordination protocol for medical-grade computation, governance, and verifiable artifacts. It enables trust-minimized medical AI infrastructure without performing diagnosis, treatment, or autonomous clinical decision-making. Built on opBNB


The Problem

Medical AI systems fail trust tests because they are: Opaque Non-reproducible Weakly governed Difficult to audit Risky to deploy legally

The Solution

NeuroGrid separates intelligence, governance, and enforcement into clean, auditable layers. No black boxes. No silent execution. No unsafe claims.


System Architecture (Simplified)

┌──────────────────────────────────────────────┐ │ USERS / JUDGES │ │ (Observers, Evaluators, DAO View) │ └──────────────────────────────────────────────┘ │ ▼ ┌──────────────────────────────────────────────┐ │ NEUROGRID – CORE (ON-CHAIN) │ │ LIVE ON opBNB TESTNET │ │ │ │ • Compute Registry │ │ • Artifact Registry │ │ • Validator Manager │ │ • Deterministic Telemetry │ │ • Fault Codes & Slashing Logic │ │ │ │ → Smart contracts enforce coordination, │ │ accountability, and auditability │ └──────────────────────────────────────────────┘ │ │ (Deterministic Signals) ▼ ┌──────────────────────────────────────────────┐ │ MEDINTEL (OFF-CHAIN – STUB) │ │ AI / Analytics / Inference Layer │ │ │ │ • No execution during hackathon │ │ • No patient data │ │ • Interfaces only │ │ │ │ → Future AI pipelines connect here │ │ under governance and audit control │ └──────────────────────────────────────────────┘ │ │ (Governance & Provenance) ▼ ┌──────────────────────────────────────────────┐ │ DESCI LAYER (SILENT / ROOT) │ │ Ethics • Provenance • Research Governance │ │ │ │ • No token logic │ │ • No execution paths │ │ • DAO & policy coordination only │ │ │ │ → Ensures long-term compliance, trust, │ │ and research integrity │ └──────────────────────────────────────────────┘


What Is Live (Hackathon-Valid)

NeuroGrid-Core smart contracts Deployed on opBNB Testnet ≥2 successful transactions Open-source No prior hackathon participation


What Is Intentionally Stubbed

AI models (MedIntel) Clinical logic Patient data handling This is by design for safety, legality, and audit clarity.


Why This Matters to BNB Chain

Strong fit for infrastructure, governance, and scalability Demonstrates responsible AI architecture Ready for future DeSci + AI ecosystems Clear upgrade path post-hackathon


Killer Use Case (Conceptual)

Verifiable Medical Analytics Pipeline External analytics run off-chain Results are signed Artifacts are registered on-chain Validators attest correctness Faults trigger slashing No diagnosis. Full accountability.


Status

Core logic: Frozen Architecture: Complete Deployment: Verified Expansion: Post-hackathon


NeuroGrid-Core (Full README)

Executive Summary

NeuroGrid is a deterministic coordination and accountability layer for medical-grade computation, research workflows, and AI-assisted analysis.

It enables: Verifiable compute execution Auditable artifact lineage Fault-coded telemetry Validator accountability DAO-governed policy enforcement NeuroGrid does not perform diagnosis or treatment. It is infrastructure, not a medical device.

This repository (neurogrid-core) contains the on-chain core protocol deployed on opBNB Testnet.


CI / Automation Status (Hackathon Scope)

Continuous Integration (CI) workflows are intentionally disabled during the hackathon submission window.

Reasoning:

  • Core protocol logic is finalized, frozen, and deployed on opBNB testnet
  • No further contract mutations are permitted pre-judging
  • CI scaffolding (extended tests, fuzzing, gas profiling) is scheduled post-hackathon

This avoids false-negative CI signals caused by:

  • intentionally stubbed MedIntel modules
  • placeholder governance and validator flows
  • environment-specific testnet configuration

CI will be re-enabled post-hackathon with:

  • full test matrix
  • coverage reports
  • security scanning
  • upgrade safety checks

What NeuroGrid Is (and Is Not)

NeuroGrid IS

A compute coordination protocol A deterministic trust framework A governance and accountability layer A compliance-aware substrate for medical AI

NeuroGrid IS NOT

A diagnostic engine A clinical decision-maker An autonomous AI doctor A patient data processor NeuroGrid supports compliant systems — it does not replace clinicians, regulators, or medical judgment.


System Architecture (Full Stack)

NeuroGrid is a three-layer system, intentionally modular.

┌────────────────────────────────────────────┐ │ DeSci Layer (Silent) │ │ Research registries • Ethics • Provenance │ │ Funding logic • DAO-aligned coordination │ │ (Off-chain, open-source, non-executing) │ └────────────────────────────────────────────┘ ▲ │ references / governance signals │ ┌────────────────────────────────────────────┐ │ MedIntel Layer (Stubbed) │ │ AI pipelines • RPM logic • Analytics │ │ NO diagnosis • NO patient data │ │ Future clinical integrations │ └────────────────────────────────────────────┘ ▲ │ deterministic execution │ ┌────────────────────────────────────────────┐ │ NeuroGrid-Core (THIS REPO) │ │ On-chain coordination & enforcement │ │ Validators • Compute • Artifacts • DAO │ │ Deployed on opBNB Testnet │ └────────────────────────────────────────────┘


Repository Scope (This Repo)

Neurogrid-core contains ONLY the on-chain protocol.

Included:

Deterministic compute coordination Validator registration and slashing Artifact issuance and lineage tracking Fault-coded telemetry enforcement Governance hooks (DAO-controlled)

Excluded by design:

AI models Medical logic Patient data Off-chain execution engines


Core On-Chain Components

Contract ### Purpose

NeuroGridKernel Global system coordinator ComputeRegistry Compute node lifecycle ArtifactRegistry Artifact issuance & lineage ValidatorManager Validator scoring & slashing NeuroDAO Governance & policy control ProposalExecutor Deterministic execution Telemetry Interface Fault-coded reporting


Deterministic Trust Model

NeuroGrid enforces trust via explicit rules, not assumptions. Deterministic execution paths Immutable fault codes Slashable validator behavior Explicit failure states

❌ No silent success ❌ No silent failure ❌ No opaque inference


Telemetry & Fault Codes

Every execution emits: Deterministic telemetry Explicit fault codes Verifiable outcomes

Key documentation:

TELEMETRY.md

FAULT_CODE.md

DETERMINISTIC_SLASHING_POLICY_SPEC.md


Token Model (Hackathon-Safe) During Hackathon

MedToken (placeholder)

Non-speculative Non-transferable utility Used for: Validator incentives Governance simulation Fault accountability

Post-Hackathon (Planned)

ABSToken

Subject to: Audit Legal review Regulatory assessment No token launch, listing, or monetization occurs during the hackathon.


Security Posture

No privileged backdoors No hidden upgrade paths Upgradeability gated by governance Explicit role separation Defense-in-depth design

See:

SECURITY_MODEL.md

STRUCTURAL_HARDENING.md

DEFAULT_ADMIN_ROLE.md

GRID_ADMIN_ROLE.md


Deployment (opBNB Testnet)

Network: opBNB Testnet Deterministic deployment order Publicly verifiable contracts ≥2 successful on-chain transactions

See:

DEPLOYMENT_BNB_TESTNET.md


Hackathon Compliance (BNB Chain)

✔ Deployed on opBNB Testnet ✔ Open source ✔ No prior hackathon participation ✔ Public GitHub repositories ✔ Live contract with transactions ✔ Clear demo and documentation ✔ Infrastructure-only (no medical claims)


Real-World Use Case (Example)

Scenario: Medical AI Research Validation Research compute is registered Execution produces verifiable artifacts Validators attest to correctness Telemetry logs faults or success DAO governs future policy updates

This enables:

Trust-minimized medical research Auditable AI workflows Regulatory-aligned infrastructure


Roadmap (Post-Hackathon)

Independent security audit Validator network expansion MedIntel controlled activation DeSci ecosystem integrations Compliant enterprise pilots


Repository Status

Core protocol: Finalized Documentation: Judge-optimized Tests: Defined & passing Deployment: Live on opBNB Testnet Upgrade state: Frozen (v1.0.0)


Final Note

NeuroGrid is intentionally conservative. It is built to earn trust before demanding adoption. Status: Core Protocol – Audit Frozen Release: v1.0.0 Network: opBNB Testnet