AetherBrain / PHDM – Mathematical Framework

Extracted from: AetherBrain Architecture: The Case for Geometric Governance and Stateful Control Layers in Large Language Models (Executive Summary, February 2026)

This document isolates the mathematical and systems-theoretic content from the AetherBrain/PHDM executive summary. No new primitives are introduced; the innovation is in composition.


1. System Decomposition

1.1 Two-Layer Intelligence Model

The system defines a two-layer computational architecture:

Layer 1 – Probabilistic Generator (Cortex)

A stochastic function:

f_theta : (x, h_t) -> y

Layer 2 – Deterministic Control Shell (Skull / Brain Case)

A stateful, rule-driven operator:

G : (s_t, y, c_t) -> {ALLOW, BLOCK, TRANSFORM}

Key property: one layer is non-deterministic (generation), one is deterministic and auditable (control). This is control theory, not ML.


2. Geometric Governance

2.1 Hyperbolic Permission Space

The system operates on the Poincare disk / ball model:

B^n = { u in R^n : ||u|| < 1 }

Hyperbolic distance d_H(u, v) serves as:

  • A partial order on permissions
  • A monotone risk scalar
  • A hierarchical embedding (tree-like structure)

2.2 Hierarchical Permissioning

Hyperbolic space is chosen because:

  • Tree depth is proportional to radial distance
  • Boundary = infinite cost
  • Near-origin = dense, low-cost region

Permission function:

Permission(s) = phi(d_H(s, s_0))

where phi is monotone. This is order theory embedded in geometry.


3. Distributed Concealment Storage

3.1 Erasure Coding

Reed-Solomon codes over finite fields:

Encode: D -> {S_1, ..., S_n}
Decode: {S_i1, ..., S_ik} -> D    (any k of n shards reconstruct)

Loss tolerance: n - k shards can be lost without data loss.

3.2 “Holographic” Property

Mathematically, this is:

  • Redundant linear reconstruction
  • Information not localized to a single address
  • Distributed linear algebra with fault tolerance

4. Reflexive Gating (Formal Control)

4.1 Deterministic Gating

Pre-execution constraint checks with hard decision boundaries:

Gate(a_t, s_t) = {
    ALLOW   if g(s_t, a_t) <= tau
    DENY    otherwise
}

This is static analysis + runtime enforcement, comparable to hardware interrupt handling.

4.2 Tokenized Authority

Capability-based access using cryptographic tokenization:

Capability = <id, scope, constraints>

Aligns with object-capability models and access control logic.


5. Control Physics

5.1 Energy Landscape

An implicit potential function:

V(s) = f(d_H(s, s_0))

Where:

  • Low energy = allowed behavior
  • High energy = forbidden behavior

Equivalent to Lyapunov-style stability analysis and barrier functions in control theory.

5.2 Timing Separation (Reflex Arcs)

Two distinct timing domains:

  • Fast deterministic layer (control) – reflex arcs
  • Slow stochastic layer (generation) – deliberation

This matches real-time systems and safety-critical control loop architecture.


6. Structural Novelty

Not new math, but new composition:

Innovation Description
A Geometry used as authorization lattice, not embedding trick
B Erasure coding used as concealment, not durability only
C Deterministic gating placed outside the probabilistic model
D Stateful control treated as physics, not policy

This is systems mathematics, not ML novelty.


7. Scope Boundaries

What this framework does not contain:

  • No new cryptographic primitives
  • No new hardness assumptions
  • No new learning algorithm
  • No new geometry theorems

The contribution is architectural composition of known mathematical structures.


8. Minimal Mathematical Identity

Compressed to one statement:

AetherBrain = deterministic control system over a hyperbolic state space, regulating a stochastic generator via geometric constraints and algebraic storage redundancy


9. Mapping to SCBE-AETHERMOORE Implementation

Mathematical Object Implementation
Poincare ball B^n src/harmonic/hyperbolic.ts
Hyperbolic distance d_H src/harmonic/hyperbolic.ts, src/harmonic/adaptiveNavigator.ts
Potential function V(s) src/harmonic/harmonicScaling.ts (Harmonic Wall, L12)
Deterministic gate G src/harmonic/pipeline14.ts (L13 risk decision)
Erasure coding src/crypto/envelope.ts, src/crypto/pqc.ts
Capability tokens src/crypto/nonceManager.ts, src/crypto/replayGuard.ts
Permission ordering src/harmonic/languesMetric.ts (Sacred Tongues 6D metric)
Energy landscape H(d,R) = phi^d / (1 + e^-R) in harmonicScaling.ts
Timing separation 14-layer pipeline (fast L1-L8 control, slow L9-L14 analysis)
PHDM drift monitor src/harmonic/phdm.ts
Quantum Axiom Mesh src/symphonic_cipher/scbe_aethermoore/axiom_grouped/

References

  • SCBE-AETHERMOORE 14-Layer Architecture: LAYER_INDEX.md
  • System Architecture: SYSTEM_ARCHITECTURE.md
  • Security Model: SECURITY.md
  • Hyperbolic geometry implementation: src/harmonic/hyperbolic.ts
  • Harmonic wall formula: src/harmonic/harmonicScaling.ts

© 2026 Aethermoore - Issac Davis, Founder | Patent Pending (63/961,403) | Products | Demo

This site uses Just the Docs, a documentation theme for Jekyll.