Maximum Entropy Principle: Ising Inference

MaxEnt: Inverse Ising

Given observed mean spins ⟨σᵢ⟩ and correlations ⟨σᵢσⱼ⟩, find the max-entropy distribution consistent with these constraints → Ising model.

max S[P] s.t. ⟨σᵢ⟩=mᵢ, ⟨σᵢσⱼ⟩=Cᵢⱼ

→ P(σ) ∝ exp(Σhᵢσᵢ + Σ Jᵢⱼσᵢσⱼ)

Target ⟨σ⟩ (mean): 0.00 Target ⟨σᵢσⱼ⟩: 0.30 Grid size N: 12 Learning rate η: 0.10

Inferred h: -
Inferred J: -
Empirical ⟨σ⟩: -
Empirical ⟨σσ⟩: -
ΔS step: -