Reference & Location

Paper: Oliva del Moral et al., Guiding extrapolations from symmetry decays for efficient error mitigation. arXiv:2603.13060 (2026).

Python module: scpn_quantum_control.mitigation.symmetry_decay

Rust acceleration: scpn_quantum_engine.fit_symmetry_decay, scpn_quantum_engine.guess_extrapolate_batch

Tests: 25 strong tests in tests/test_symmetry_decay.py

1. The Problem Standard ZNE Does Not Solve

Richardson zero-noise extrapolation fits a polynomial through observables measured at amplified noise levels (via circuit folding) and extrapolates to the zero-noise limit. At large circuit depths (100+ qubits, thousands of CZ gates), the polynomial model fails to capture the true noise profile, leading to systematic over- or under-correction. Oliva del Moral et al. document divergence of Richardson ZNE beyond depth 2000 CZ gates on IBM hardware (their Figure 7).

Polynomial extrapolation assumes noise is additive and scale-invariant, which breaks down for correlated noise channels. No standard method leverages physical conservation laws specific to the Hamiltonian being simulated.

2. The Physics Insight

If the Hamiltonian $H$ conserves a symmetry observable $S$ (i.e. $[H, S] = 0$), then $\langle S \rangle$ is analytically known for any initial state. The deviation of $\langle S \rangle$ from its ideal value under hardware noise directly reveals the noise-induced decay profile.

GUESS transfers this learned decay to target observables whose ideal values are unknown. Instead of fitting a generic polynomial, GUESS uses physics to constrain the extrapolation.

3. Why the XY Hamiltonian Gets It Free

The SCPN Kuramoto-XY Hamiltonian

$$H_{XY} = \sum_{n, m} K_{nm} \left( \sigma_n^x \sigma_m^x + \sigma_n^y \sigma_m^y \right) + \sum_i \omega_i Z_i$$

naturally conserves total magnetisation $S = \sum_i Z_i$ because the $XX + YY$ interaction flips pairs of spins in opposite directions and leaves the total $Z$-component invariant. Formally $[H_{XY}, \sum_i Z_i] = 0$.

This is the crucial point: measuring $S = \sum_i Z_i$ requires only Z-basis measurements, which are already part of every experiment run. Whereas generic ZNE costs an additional 2–4× shot budget per noise scale factor, GUESS on the XY Hamiltonian costs zero extra shots. The symmetry observable is a free rider on the target observable measurement.

4. Mathematical Formulation

4.1 Exponential decay model

Under noise at scale factor $g$ (where $g = 1$ is base noise), the symmetry observable decays exponentially:

$$\langle S \rangle_g = \langle S \rangle_\text{ideal} \cdot e^{-\alpha (g - 1)}$$

where $\alpha \ge 0$ is the noise scaling exponent. This model follows from the Lindblad master equation under depolarising noise: each gate contributes an independent decay factor, and circuit folding multiplies the total decay rate by $g$.

4.2 Learning $\alpha$ via log-linear regression

Taking the logarithm:

$$\ln \left( \frac{\langle S \rangle_g}{\langle S \rangle_\text{ideal}} \right) = -\alpha (g - 1)$$

This is a linear model $y = -\alpha \cdot x$ where $y_i = \ln(\langle S \rangle_{g_i} / \langle S \rangle_\text{ideal})$ and $x_i = g_i - 1$. We fit $\alpha$ via ordinary least-squares on $(x_i, y_i)$ pairs from $N \ge 2$ noise scale measurements.

Fit residual: $r = \sqrt{N^{-1} \sum_i (y_i - \hat{y}_i)^2}$. A large residual ($r > 0.1$) flags non-Markovian noise or systematic calibration drift and triggers a fallback to raw values.

4.3 GUESS extrapolation formula

Given the learned $\alpha$, the mitigated value of any target observable $O$ is (Oliva del Moral et al., 2026, Eq. 5):

$$\langle O \rangle_\text{mitigated} \;\approx\; \langle O \rangle_\text{noisy} \cdot \left( \frac{\left|\langle S \rangle_\text{ideal}\right|}{\left|\langle S \rangle_\text{noisy}\right|} \right)^\alpha$$

Properties. When noise is absent, $C = 1$ and no correction is applied. When $\alpha = 0$, $C = 1$ regardless of symmetry values. When the symmetry fully decays ($\langle S \rangle_\text{noisy} \to 0$), the correction diverges and we fall back to the raw value. $C \ge 1$ for physical noise.

5. Phase 1 Noise Profile Calibration

The ibm_kingston Phase 1 campaign (342 circuits, 8 Trotter depths) provides exactly the noise profile GUESS needs. Parity leakage rises smoothly from ~8 % at depth 2 to ~28 % at depth 30 — a textbook exponential-plus-saturation curve.

For the Phase 2 campaign, a noise-scaled sub-sweep with circuit folding factors $g \in \{1, 3, 5\}$ will use the measured parity leakage itself as the symmetry observable. Because parity leakage is the complement of $\langle P \rangle = \langle \prod_i Z_i \rangle$, this is mathematically equivalent to running GUESS on the $\mathbb{Z}_2$ parity operator, with the ideal value $\langle P \rangle_\text{ideal} = \pm 1$ known in closed form.

→ See the Phase 1 results page for the full leakage-vs-depth curve.

6. Code Example

# Full GUESS pipeline on the XY Hamiltonian — end-to-end
from scpn_quantum_control.mitigation.symmetry_decay import GUESSMitigator
from scpn_quantum_control.bridge import KnmCompiler
from scpn_quantum_control.phase import TrotterEvolution
import numpy as np

# 1. Build the XY Hamiltonian (4 oscillators, exp-decay coupling)
n = 4
K = 0.45 * np.exp(-0.3 * np.abs(np.subtract.outer(range(n), range(n))))
np.fill_diagonal(K, 0.0)
omega = np.linspace(0.8, 1.2, n)
H = KnmCompiler(K, omega).hamiltonian()

# 2. Build the Trotter circuit at depth d=10
circuit = TrotterEvolution(H, dt=0.3, steps=10).circuit()
circuit.measure_all()

# 3. Run at three noise scales (circuit folding factors 1, 3, 5)
mitigator = GUESSMitigator(symmetry="total_magnetisation")
result = mitigator.run(circuit, backend="ibm_kingston", scales=[1.0, 3.0, 5.0])

print(f"Learned alpha: {result.alpha:.3f}")
print(f"Fit residual: {result.residual:.4f}")
print(f"Raw target (no mitig.): {result.raw_target:.4f}")
print(f"GUESS-mitigated target: {result.mitigated_target:.4f}")

The symmetry observable $\langle S \rangle$ and the target observable $\langle O \rangle$ are extracted from the same Z-basis counts. No extra shots are required to measure $S$; it is a post-processing step on the same bitstrings.

7. Comparison with Alternative Mitigation Methods

MethodPhysics-awareShot overheadWorks at deep circuitsBest for
GUESS (this page)0 (for XY)Hamiltonians with known symmetry
Richardson ZNE2–4×diverges > 2000 CZgeneric circuits, shallow depth
Exponential ZNE2–4×generic circuits, deep depth
PECnoise-modelexp. in depthhighest accuracy when noise model is known
DDD0 (idle slots only)idle-qubit low-frequency noise
$\mathbb{Z}_2$ parity post-selectionpartial (shots rejected)hard discard of bad shots

GUESS and DDD are complementary and can be stacked. GUESS and post-selection target the same symmetry; GUESS extrapolates while post-selection discards. They can also be stacked.

8. Limitations

When GUESS will not help you

  • No symmetry. If your Hamiltonian has no easily-measurable conserved observable, GUESS is not applicable. For circuits with no known symmetry, use Richardson/exponential ZNE or PEC.
  • Non-Markovian noise. The exponential decay model breaks down if the noise is strongly non-Markovian. A large fit residual ($r > 0.1$) will flag this and fall back to raw values.
  • Symmetry-breaking coupling. If the experiment contains gates that break the symmetry (e.g. stray $X$ rotations), the decay profile will be contaminated by coherent effects, not just decoherence.
  • Full decay. If the circuit is so deep that $\langle S \rangle_\text{noisy} \to 0$, the correction factor diverges and GUESS falls back to the raw value. This is the right behaviour but means no mitigation is possible in that regime.

Related Pages

The Phase 1 ibm_kingston noise profile that calibrates GUESS for Phase 2.
The parity symmetry that GUESS exploits — proof and hardware confirmation.
Full algorithm catalogue including alternative mitigation methods.