Adaptive Policy-Guided Error Mitigation (APGEM): A Unified Learning-Based Framework for Noise Resilience in Quantum Machine Learning
Organizations: Department of Information Technology National Institute of Technology Srinagar Jammu & Kashmir, India
Abstract
Error mitigation stands between variational quantum algorithms and useful results on today's noisy hardware, yet in practice a mitigation method is usually chosen once, by hand, and applied unchanged for the life of an experiment. This is a poor fit for quantum machine learning (QML), where the circuit and the noise drift as training proceeds, so the best strategy in one regime is often wrong in another. We present a generalized form of Adaptive Policy-Guided Error Mitigation (APGEM), which treats the choice of mitigation technique as an online decision made from an algorithm's own feedback. A single contextual-bandit controller observes signals any variational algorithm already produces, such as fidelity, entropy, reward or loss, and gradient behavior, and selects among Zero-Noise Extrapolation, Probabilistic Error Cancellation, Measurement Error Mitigation, Dynamical Decoupling, and circuit optimization per step. Because it depends only on this shared feedback interface, the framework applies unchanged across supervised, unsupervised, reinforcement, and hybrid quantum-classical learning. Evaluated on five paradigms across 80 algorithm-and-noise conditions, adaptive selection reaches 95.5% of an oracle that always picks the best technique and significantly outperforms all six fixed strategies (Friedman , ; pairwise Wilcoxon ). The learned policy is noise-aware: its favored technique shifts as depolarizing and readout error rise. On noise models from eight real IBM devices, APGEM tracks the oracle almost exactly (99.9% of oracle utility) at a flat cost of roughly shots from three to eight qubits.
Figures & tables
| Ref | Key Contribution | Method Used | Target Problem | Noises Addressed | Error Mitigation Used and Focus |
|---|---|---|---|---|---|
| [ 23 ] | Systematic ML-QEM study showing classical models can emulate conventional mitigation at lower runtime overhead. | Supervised models (linear regression, random forests, MLPs, graph neural networks) predicting mitigated observables. | Scalable, low-overhead mitigation across diverse circuits and device-noise profiles. | Device-level stochastic noise (interpolation and extrapolation). | ML-QEM; reduces sampling overhead by mimicking traditional QEM. |
| [ 24 ] | ML-based mitigation framed as a reliable quantum software-development workflow on IBM hardware. | Machine-learning correction of measured outputs, evaluated as a software-engineering pipeline. | Dependable execution of quantum programs under device noise. | Hardware noise on IBM superconducting devices. | Learning-based QEM from a software-reliability perspective. |
| [ 25 ] | Adaptive neural network that dynamically adjusts expectation values from error characteristics on a 127-qubit device. | Neural network mapping noisy to corrected observables, adapting to device error profiles. | Dynamic, device-aware mitigation on large superconducting processors. | Time- and device-dependent hardware noise. | Adaptive learning-based QEM; neural analogue to policy-guided adaptation. |
| [ 26 ] | Practical ML-QEM for variational algorithms with per-task model selection and transfer across Hamiltonians. | Regression models (Ridge, XGBoost) trained on simulated near-Clifford data and selected per task. | Parameter- and Hamiltonian-transferable mitigation for VQAs, benchmarked on VQE. | Circuit- and depth-dependent noise, high-noise regime. | ML-QEM with model selection; conceptually aligned with adaptive selection. |
| [ 15 ] | Foundational framework introducing zero-noise extrapolation (ZNE) and probabilistic error cancellation (PEC) for NISQ devices. | Analytical extrapolation and quasi-probability correction. | Bias reduction in short-depth quantum circuits. | Gate and readout noise (Markovian). | ZNE and PEC; analytical and hardware-compatible, but sampling-intensive. |
| [ 16 ] | Learning-based QEM trained on near-Clifford circuits, removing the need for an explicit noise model. | Data-driven regression on noisy–ideal circuit pairs. | Hardware noise suppression without explicit noise models. | Device-specific stochastic and coherent noise. | Learning-based QEM; emphasises adaptability and noise-model independence. |
| Adapter | Paradigm | Native indicator | Normalization | |
|---|---|---|---|---|
| VQC | supervised | parity-readout classification accuracy | identity on | 0.75 |
| QSVM | supervised (kernel) | test accuracy of an SVM on the quantum kernel | identity on | 0.75 |
| VQE | hybrid eigensolver | energy of a transverse-field Ising chain | 0.80 | |
| QGen | unsupervised generative | total-variation distance to the target | 0.70 | |
| QRL | reinforcement | tour cost on a seven-node routing MDP | , clipped to | 0.60 |
| Technique | Shots per application | Overhead |
|---|---|---|
| NoMitigation | 6,554 | |
| DD | 6,554 | |
| CircuitOpt | 6,554 | |
| MEM | 13,107 | |
| ZNE | 19,661 | |
| PEC | 26,214 |
| Qubits | Depth | Parameters | Shots per decision | Wall time (s) |
|---|---|---|---|---|
| 3 | 7 | 6 | 13,312 | 5.1 |
| 4 | 9 | 8 | 13,312 | 2.2 |
| 5 | 11 | 10 | 13,312 | 2.2 |
| 6 | 13 | 12 | 13,312 | 2.4 |
| 7 | 15 | 14 | 13,312 | 2.6 |
| 8 | 17 | 16 | 13,312 | 2.5 |
| Readout error | Two-qubit gate error | ( s) | ( s) | |
|---|---|---|---|---|
| Range across 8 devices | 0.014–0.037 | 0.006–0.013 | 62–148 | 39–115 |
| Component | Version |
|---|---|
| Python | 3.10.11 |
| Qiskit | 1.4.0 |
| Qiskit Aer | 0.16.1 |
| NumPy | 2.2.6 |
| SciPy | 1.14.1 |
| scikit-learn | 1.7.1 |
| Method | Mean utility | CI | Oracle match | Mean regret |
|---|---|---|---|---|
| APGEM (LinUCB) | 0.435 | [0.323, 0.542] | 0.558 | 0.0755 |
| Thompson sampling | 0.431 | [0.315, 0.541] | 0.533 | 0.0587 |
| -greedy | 0.398 | [0.280, 0.508] | 0.508 | 0.0761 |
| Static NoMitigation | 0.356 | [0.237, 0.477] | 0.400 | 0.1037 |
| Static PEC | 0.343 | [0.200, 0.468] | 0.263 | 0.1125 |
| Static MEM | 0.337 | [0.221, 0.454] | 0.087 | 0.1227 |
| Method | Avg. rank | Wilcoxon | Cliff’s |
|---|---|---|---|
| APGEM | 2.375 | — | — |
| NoMitigation | 3.013 | 0.101 | |
| CircuitOpt | 4.287 | 0.176 | |
| PEC | 4.431 | 0.156 | |
| MEM | 4.481 | 0.172 | |
| DD | 4.619 | 0.192 |
| Paradigm | Dominant choice (share) | APGEM util. | Oracle util. | Argmax match |
|---|---|---|---|---|
| VQC | DD / PEC (0.25 each) | 0.465 | 0.497 | 0.38 |
| QSVM | NoMitigation (0.62) | 0.078 | 0.095 | 0.69 |
| VQE | PEC (0.56) | 0.370 | 0.375 | 0.75 |
| QGen | PEC (0.62) | 0.526 | 0.530 | 0.81 |
| QRL | NoMitigation (0.38) | 0.743 | 0.790 | 0.25 |
| Noise regime | Blocks | APGEM top (share) | Oracle top (share) |
|---|---|---|---|
| Very low | 5 | NoMitigation (0.40) | NoMitigation (1.00) |
| Low | 5 | ZNE (0.40) | ZNE (0.60) |
| Moderate | 20 | PEC (0.60) | PEC (0.55) |
| Severe | 20 | NoMitigation (0.40) | NoMitigation (0.55) |
| Readout-dominated | 10 | NoMitigation (0.30) | NoMitigation (0.40) |
| Coherent-dominated | 20 | PEC (0.30) | NoMitigation (0.35) |
| VQE | QGen | QSVM | VQC | QRL | |
| Noise floor | 0.0006 | 0.0076 | 0.0099 | 0.0586 | 0.1116 |
| Technique | Quality [95% CI] | Fidelity | Entropy | Shots | Overhead | HW (s) |
|---|---|---|---|---|---|---|
| PEC | 0.702 [0.666, 0.738] | 0.741 | 0.777 | 26214 | 10.49 | |
| ZNE | 0.645 [0.611, 0.681] | 0.692 | 0.864 | 19661 | 7.86 | |
| MEM | 0.627 [0.593, 0.661] | 0.678 | 0.907 | 13107 | 5.24 | |
| NoMitigation | 0.618 [0.585, 0.651] | 0.661 | 0.928 | 6554 | 2.62 | |
| CircuitOpt | 0.589 [0.556, 0.622] | 0.642 | 0.939 | 6554 | 2.62 | |
| DD | 0.583 [0.550, 0.616] | 0.637 | 0.944 | 6554 | 2.62 |
| Depol. rate | APGEM | NoMitigation | Best possible |
|---|---|---|---|
| 0.01 | 0.838 | 0.777 | 0.858 |
| 0.05 | 0.749 | 0.659 | 0.766 |
| 0.12 | 0.660 | 0.540 | 0.675 |
| 0.18 | 0.534 | 0.495 | 0.598 |
| Qubits | Depth | 2Q gates | Params | Decision shots | Best arm | Quality | Wall (s) |
|---|---|---|---|---|---|---|---|
| 3 | 7 | 4 | 6 | 13312 | PEC | 0.893 | 5.06 |
| 4 | 9 | 6 | 8 | 13312 | PEC | 0.872 | 2.19 |
| 5 | 11 | 8 | 10 | 13312 | PEC | 0.885 | 2.24 |
| 6 | 13 | 10 | 12 | 13312 | PEC | 0.875 | 2.41 |
| 7 | 15 | 12 | 14 | 13312 | PEC | 0.862 | 2.58 |
| 8 | 17 | 14 | 16 | 13312 | PEC | 0.810 | 2.47 |
| Horizon | Mean utility [95% CI] | Oracle match | Mean regret | Total shots |
|---|---|---|---|---|
| 100 | 0.420 [0.309, 0.527] | 0.500 | 0.1117 | |
| 400 | 0.435 [0.323, 0.542] | 0.558 | 0.0755 | |
| 1000 | 0.440 [0.327, 0.546] | 0.550 | 0.0525 |
| Readout error | Two-qubit error | ( s) | ( s) | |
|---|---|---|---|---|
| Range across 8 devices | 0.014–0.037 | 0.006–0.013 | 62–148 | 39–115 |
| Quantity | Value |
|---|---|
| Blocks | 32 |
| APGEM utility (% of oracle) | 99.9% |
| Exact argmax match | 0.594 |
| Within-tolerance agreement ( ) | 1.000 |
| Friedman | 128.68 |
| Friedman |
Quantum Error Mitigation
Quantum error mitigation estimates more reliable circuit outputs on noisy processors without requiring full fault-tolerant quantum computing.
See all 13 papers