Back to Amplified Intelligence
Ψ₂₄ Hilbert State Engine
A 24-dimensional Hilbert space for market psychology. Tiers 1–4 span behavioral dynamics, narrative mechanics, quantum microstructure, and non-linear thermodynamics. Evolved on-device via WebGPU; exotic dimensions estimated by a grounded LLM.
Backend: cpuTicks: 0
|ψₜ⟩ projection — all 24 dimensions
- Classical Behavioral
- Narrative & Structural
- Quantum Microstructure
- Non-Linear Thermodynamics
Exotic dims (LLM)
Start the engine to fetch R, ξ, D, Φ from grounded LLM search.
Regime Classifier
Awaiting Ψ evolution…
Sentiment vs Regime — Live Divergence
Start the engine to pull live news + social chatter for NVDA.
Ψ₂₄ component breakdown
Tier 1 · Classical Behavioral
Surface-level emotional drivers — retail + institutional price action.
Hₜ 0.000
Hype
Fₜ 0.000
Fear
Cₜ 0.000
Conviction
Sₜ 0.000
Surprise
Gₜ 0.000
Greed
Xₜ 0.000
Exhaustion
Pₜ 0.000
Polarization
Mₜ 0.000
Macro Trust
Tier 2 · Narrative & Structural
How information and liquidity physically travel through the market graph.
Eₜ 0.000
Info Entropy
Vₜ 0.000
Contagion
Rₜ ✦0.000
Reflexivity
Lₜ 0.000
Viscosity
Tier 3 · Quantum Microstructure
Hidden non-linear forces in dark pools and algorithmic sub-routines.
ξₜ ✦0.000
Entanglement
Wₜ 0.000
Superposition
Dₜ ✦0.000
Dark Mass
τₜ 0.000
Causal Lag
Tier 4 · Non-Linear Thermodynamics
Kinetic limits and permanent structural evolution over time.
Iₜ 0.000
Inertia
Aₜ 0.000
Asymmetry
Oₜ 0.000
Oscillation
Kₜ 0.000
Kinematics
Zₜ 0.000
Zero-Point
Bₜ 0.000
Bifurcation
Yₜ 0.000
Hysteresis
Φₜ ✦0.000
Phase Coherence
Architecture · Phases A · B · C · D
- Phase A — Numerics: Complex-valued ψ ∈ ℂ²⁴, Schrödinger RK4 with Hermitian H (imaginary Tier-3/4 couplings → genuine interference), Lindblad decoherence on Eₜ, Yₜ, Xₜ.
- Phase B — Latency: Persistent GPU buffer pool (no per-tick allocs), ping-pong A/B buffers, async readback. Batched multi-ticker kernel available via
HilbertBatchKernel. - Phase C — Signal: Live yfinance features stream into Tier-1/4 drive every 60s. LLM exotic dims passed through a 1-D Kalman filter (Q = 0.005, R = (1−conf)²·0.5). 3-model ensemble (Gemini + Claude + GPT) available via toggle.
- Phase D — Actionability: 6-state regime classifier (melt-up / distribution / accumulation / capitulation / ranging / breaking-down). Φ alert toast when |Φ| ≥ 0.65. Auto-persists
HilbertSnapshotevery 50 ticks. Backtest harness scores snapshots against realized forward returns. - Tier 1 seed: 6-emotion vector from EmotionTensorForge mapped to {H,F,C,S,G,X,P,M}.
Computation on the edge. Coherence preserved.