structure-aware-variance-reduction-hamiltonian

name	structure-aware-variance-reduction-hamiltonian
description	Variance reduction methodology for unbiased randomized Hamiltonian simulation. Applies classical variance reduction techniques to randomized product formulas (continuous TE-PAI) without introducing bias, achieving discretization-error-free simulation with finite-depth random circuits. Use when optimizing quantum Hamiltonian simulation, reducing Trotter error, or analyzing variance-sampling tradeoffs in randomized quantum algorithms.
metadata	{"arxiv_id":"2606.23544","published":"2026-06-22","authors":"Joshua W. Dai, Fredrik Hasselgren, Chusei Kiumi","tags":["hamiltonian-simulation","variance-reduction","randomized-algorithms","trotter-error","quantum-simulation"]}

Structure-Aware Variance Reduction for Hamiltonian Simulation

Core Concepts

Randomized Hamiltonian simulation trades systematic bias for sampling overhead. This framework applies classical variance reduction to unbiased randomized protocols (continuous TE-PAI) without introducing additional bias, dramatically reducing the number of samples needed.

Key Innovation

Continuous TE-PAI (Time-Evolution Probabilistic Angle Interpolation): removes Trotter discretization error with finite-depth random circuits, whereas deterministic Trotterization requires infinite depth.

Critical Finding

Discretization error in Trotterized simulations causes unphysical exponential growth in bond dimension for tensor-network simulations. Continuous TE-PAI at comparable depth avoids this growth entirely.

Methodology

Step 1: Continuous TE-PAI Protocol

For Hamiltonian H = Σⱼ Hⱼ:

1. Sample random circuit depth k from geometric-like distribution
2. At each step, randomly select term Hⱼ with probability pⱼ
3. Apply exp(-iθHⱼ) with angle θ drawn from quasiprobability distribution
4. Weight output by importance sampling factor w(k, {jₘ})

Step 2: Structure-Aware Variance Reduction

Exploit Hamiltonian structure to reduce Monte Carlo variance:

• Control variates: Use deterministic Trotter as correlated control
• Stratified sampling: Partition circuit space by structural features
• Importance weighting: Preferentially sample high-contribution circuits

Step 3: Optimal Sampling Distribution

Find pⱼ that minimizes Var[estimator] subject to unbiasedness:

• Optimal pⱼ ∝ ||Hⱼ|| · contribution factor
• Requires classical precomputation of term norms and commutators

Usage Patterns

Pattern 1: Trotter-Free Simulation

When high-precision Hamiltonian simulation is needed:

• Replace deterministic Trotter with continuous TE-PAI
• Apply variance reduction to cut sample cost
• Achieves O(ε⁻¹) scaling vs O(ε⁻¹⁻¹/p) for Trotter

Pattern 2: Tensor Network Simulation

When simulating quantum circuits with tensor networks:

• Use continuous TE-PAI to avoid bond dimension explosion
• Critical for long-time evolution of 1D/2D systems

Pattern 3: Quantum Chemistry

For molecular Hamiltonian simulation:

• Group Pauli terms by commuting structure
• Apply stratified sampling within each group
• Leverage locality to reduce effective term count

Pitfalls

• Quasiprobability overhead: negative quasiprobabilities increase sampling cost — bound by L1 norm of coefficients
• Classical preprocessing: optimal sampling distribution requires term norm computation
• Not for biased protocols: variance reduction must preserve the mean channel exactly

Activation Keywords

• variance reduction Hamiltonian simulation
• continuous TE-PAI
• randomized product formula quantum
• Trotter error mitigation
• unbiased quantum simulation
• 量子哈密顿模拟方差缩减
• structure aware variance reduction