| name | particle-preserving-fermionic-shadows |
| description | Classical shadow estimation for fermionic states with mode-independent sample complexity. Improves worst-case bound for Slater determinant overlap estimation from O(√n log n) to O(η log η), using harmonic analysis on AIII symmetric space. |
Particle-Preserving Fermionic Shadows
Description
Methodology for learning expectation values of particle-preserving operators with respect to unknown η-particle n-mode fermionic states via classical shadows. Achieves mode-independent sample complexity O(η log η) for Slater determinant overlap estimation, reducing from the previous worst-case O(√n log n) bound. Uses harmonic analysis on the AIII symmetric space U(n)/(U(η) × U(n-η)) and Jacobi ensemble techniques.
Activation Keywords
- fermionic classical shadows
- mode-independent sample complexity
- Slater determinant overlap
- AIII symmetric space quantum
- Jacobi ensembles quantum
- particle-preserving operators
- fermionic state tomography
- 费米子经典阴影
- 模式无关采样复杂度
- Slater行列态重叠
Core Concepts
Classical Shadow Estimation for Fermions
- Problem: Learn expectation values of particle-preserving operators from unknown fermionic states
- Key insight: Particle-preserving structure enables mode-independent sample complexity
- Randomization: Approximate unitary designs in first-quantized encoding achieve polylogarithmic circuit depth
Sample Complexity Improvement
- Previous worst case: O(√n log n) samples
- New bound: O(η log η) samples (independent of total mode count n)
- For quadratic observables: O(η ||h₀||₂²) where h₀ is traceless component
- Classical post-processing: O(n η²) for generic dense orbital, O(n² η) for quadratic observables
AIII Symmetric Space Analysis
- Reduction: Extremal shadow variance → harmonic analysis on U(n)/(U(η) × U(n-η))
- Techniques: Jacobi ensembles and orthogonal polynomials for integral evaluation
- Significance: The mathematical technique may be of independent interest beyond shadows
Usage Patterns
Pattern 1: Fermionic State Learning
- Identify particle number η and mode count n
- Select appropriate randomization scheme (first-quantized vs second-quantized)
- Apply classical shadow protocol with O(η log η) samples
- Post-process classically in O(n η²) time
Pattern 2: Slater Determinant Overlap Estimation
- Target state: unknown η-particle n-mode fermionic state
- Reference state: arbitrary Slater determinant
- Sample complexity: O(η log η) for fixed additive precision
- Advantage: exponential improvement when η ≪ n
Mathematical Framework
Symmetric Space Decomposition
Shadow variance extremal problem
↓ reduction
Harmonic analysis on AIII symmetric space U(n)/(U(η) × U(n-η))
↓ evaluation
Jacobi ensemble integrals + orthogonal polynomials
↓ result
O(η log η) sample complexity bound
Circuit Depth Comparison
- First-quantized: Polylogarithmic depth (approximate unitary designs)
- Second-quantized matchgate: Linear depth (nearest-neighbor)
Error Handling
Sample Complexity Bounds
- If η ≈ n: The O(η log η) bound approaches O(n log n), still better than O(√n log n) for large n
- If observables not particle-preserving: Different shadow protocol required
- If classical post-processing too slow: Consider sparse orbital structure optimization
References
- arXiv:2606.27254 - Particle-preserving fermionic shadows (West, Cerezo, Larocca 2026)
- Classical shadow tomography literature
- Harmonic analysis on symmetric spaces