Machine Unlearning in the Era of Quantum Machine Learning: An Empirical Study

2025-12-22

AI Summary

Machine unlearning (MU) in hybrid quantum-classical neural networks—particularly variational quantum circuits (VQCs)—remains unexplored, with no empirical studies or benchmarks to date. Method: We propose two quantum-adapted unlearning strategies, extending classical approaches—including gradient-based updates, knowledge distillation, regularization, and certified unlearning—to the quantum regime, and design quantum-aware algorithms. Contribution/Results: Evaluated on Iris, MNIST, and Fashion-MNIST for subset removal and full-class deletion, our study reveals intrinsic forgetting stability in shallow VQCs. EU-k, LCA, and Certified Unlearning achieve optimal trade-offs among utility retention, forgetting strength, and retraining alignment. We establish the first quantum MU benchmark, identifying critical influences of circuit depth, entanglement topology, and task complexity on unlearning efficacy. All code is open-sourced to enable reproducible research in quantum privacy-preserving learning.

Technology Category

• Machine Learning: Quantum Machine Learning
• Game Theory and Economic Paradigms: Adversarial Learning
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Security and Privacy: Security and privacy of machine learning and AI applications
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

We present the first comprehensive empirical study of machine unlearning (MU) in hybrid quantum-classical neural networks. While MU has been extensively explored in classical deep learning, its behavior within variational quantum circuits (VQCs) and quantum-augmented architectures remains largely unexplored. First, we adapt a broad suite of unlearning methods to quantum settings, including gradient-based, distillation-based, regularization-based and certified techniques. Second, we introduce two new unlearning strategies tailored to hybrid models. Experiments across Iris, MNIST, and Fashion-MNIST, under both subset removal and full-class deletion, reveal that quantum models can support effective unlearning, but outcomes depend strongly on circuit depth, entanglement structure, and task complexity. Shallow VQCs display high intrinsic stability with minimal memorization, whereas deeper hybrid models exhibit stronger trade-offs between utility, forgetting strength, and alignment with retrain oracle. We find that certain methods, e.g. EU-k, LCA, and Certified Unlearning, consistently provide the best balance across metrics. These findings establish baseline empirical insights into quantum machine unlearning and highlight the need for quantum-aware algorithms and theoretical guarantees, as quantum machine learning systems continue to expand in scale and capability. We publicly release our code at: https://github.com/CrivoiCarla/HQML.

Problem

• Adapting classical unlearning methods to quantum machine learning settings
• Evaluating unlearning performance in hybrid quantum-classical neural networks
• Exploring quantum-specific factors affecting forgetting and utility trade-offs

Innovation

• Adapting classical unlearning methods to quantum settings
• Introducing two new unlearning strategies for hybrid models
• Evaluating unlearning across quantum circuit depths and tasks