Quantum machine learning (QML) models introduce unique sources of uncertainty arising from finite measurement shots, hardware noise, and variational optimization instability. Despite rapid progress in quantum algorithms, systematic uncertainty quantification (UQ) in QML remains largely underexplored and lacks standardized tooling. We present QuantumUQ, an open-source Python library for uncertainty quantification in quantum machine learning with native support for PennyLane and Qiskit. QuantumUQ provides shot based uncertainty estimation, ensemble-based epistemic uncertainty, measurement-budget stability profiling, and calibration metrics for classification and regression tasks. We formalize the uncertainty decomposition in QML, describe the architectural design of the library, and demonstrate its capabilities through reproducible experiments. Our results show that measurement induced uncertainty exhibits predictable scaling behavior and that calibration analysis reveals measurable calibration bias in the evaluated variational QML classifier.
Quantum machine learning (QML) models introduce unique sources of uncertainty arising from finite measurement shots, hardware noise, and variational optimization instability. Despite rapid progress in quantum algorithms, systematic uncertainty quantification (UQ) in QML remains largely underexplored and lacks standardized tooling. We present QuantumUQ, an open-source Python library for uncertainty quantification in quantum machine learning with native support for PennyLane and Qiskit. QuantumUQ provides shot based uncertainty estimation, ensemble-based epistemic uncertainty, measurement-budget stability profiling, and calibration metrics for classification and regression tasks. We formalize the uncertainty decomposition in QML, describe the architectural design of the library, and demonstrate its capabilities through reproducible experiments. Our results show that measurement induced uncertainty exhibits predictable scaling behavior and that calibration analysis reveals measurable calibration bias in the evaluated variational QML classifier.
Quantum machine learning (QML) is a promising paradigm for tackling computational problems that challenge classical AI. Yet, the inherent probabilistic behavior of quantum mechanics, device noise in NISQ hardware, and hybrid quantum-classical execution pipelines introduce new risks that prevent reliable deployment of QML in real-world, safety-critical settings. This research offers a broad roadmap for Trustworthy Quantum Machine Learning (TQML), integrating three foundational pillars of reliability: (i) uncertainty quantification for calibrated and risk-aware decision making, (ii) adversarial robustness against classical and quantum-native threat models, and (iii) privacy preservation in distributed and delegated quantum learning scenarios. We formalize quantum-specific trust metrics grounded in quantum information theory, including a variance-based decomposition of predictive uncertainty, trace-distance-bounded robustness, and differential privacy for hybrid learning channels. To demonstrate feasibility on current NISQ devices, we validate a unified trust assessment pipeline on parameterized quantum classifiers, uncovering correlations between uncertainty and prediction risk, an asymmetry in attack vulnerability between classical and quantum state perturbations, and privacy-utility trade-offs driven by shot noise and quantum channel noise. This roadmap seeks to define trustworthiness as a first-class design objective for quantum AI.
Quantum machine learning is a promising programming paradigm for the optimization of quantum algorithms in the current era of noisy intermediate scale quantum (NISQ) computers. A fundamental challenge in quantum machine learning is generalization, as the designer targets performance under testing conditions, while having access only to limited training data. Existing generalization analyses, while identifying important general trends and scaling laws, cannot be used to assign reliable and informative “error bars” to the decisions made by quantum models. In this article, we propose a general methodology that can reliably quantify the uncertainty of quantum models, irrespective of the amount of training data, of the number of shots, of the ansatz, of the training algorithm, and of the presence of quantum hardware noise. The approach, which builds on probabilistic conformal prediction, turns an arbitrary, possibly small, number of shots from a pre-trained quantum model into a set prediction, e.g., an interval, thatprovablycontains the true target with any desired coverage level. Experimental results confirm the theoretical calibration guarantees of the proposed framework, referred to asquantum conformal prediction.
Quantum Machine Learning (QML) is a discipline of research at the intersection of quantum information and machine learning that leverages quantum mechanical properties to enhance computational capabilities. With its emergence, there is a need to integrate QML models into machine learning pipelines for real-life applications such as image processing. While standalone programs exist to demonstrate the performance of QML models, a well-defined model workflow is noticeably absent. This thesis thoroughly explores various existing QML models and their practical utility in image processing tasks, with the aim of constructing a robust QML library. Throughout this thesis, we develop piQture, an open-source Python and Qiskit-based library that streamlines the development, training, and evaluation of QML models. Its design and structure prioritize usability among users familiar with classical machine learning without prior QML experience. Further, piQture is augmented with automated building, testing, and packaging workflows that enhance software reliability and reproducibility. Finally, we provide strategies to facilitate model management and storage within piQture for practical adoption and future analysis of pre-trained QML models.
Advancements in quantum computing have sparked a systematic evaluation of quantum machine learning (QML) models, focusing on their behavior and performance across diverse learning tasks such as classification, regression, generative modeling, and feature compression. The primary focus is on assessing the performance and behavior of various QML architectures-including Quantum Neural Networks (QNNs), Quantum Support Vector Classifiers (QSVCs), Quantum Generative Adversarial Networks (QGANs), Quantum Convolutional Neural Networks (QCNNs), and Quantum Autoencoders (QAEs)-implemented using Qiskit's machine learning library. This evaluation is conducted using an existing benchmarking framework that ensures consistent experimental conditions and supports flexible evaluation strategies. To provide additional context and insight, these quantum models are compared against classical counterparts using widely adopted implementations. Experiments are carried out on a range of datasets, including structured and low-dimensional examples like Bars and Stripes, as well as more complex datasets like Downscaled MNIST and synthetic Hyperplanes. By providing a standardized evaluation of these models within the Qiskit framework, this study establishes a reproducible performance baseline and highlights key challenges for QML systems in the noisy intermediate-scale quantum (NISQ) era.
We present an uncertainty quantification (UQ) framework for application level benchmarking and characterization of noisy quantum backends. The framework compares two workload classes under one statistical pipeline: noisy intermediate scale quantum (NISQ) variational quantum algorithms (VQAs) and Quantum Singular Value Transformation (QSVT) based Green's function reconstruction. For the VQA branch, we evaluate ten benchmark families spanning chemistry, optimization, simulation, compiling, linear solving, partial differential equations, metrology, error correction, tomography, and channel fidelity estimation. For the QSVT branch, we reconstruct orbital resolved Green's functions and spectral peaks from a block encoded real time propagator. The workflow combines Bayesian optimization, posterior distribution refinement, sensitivity analysis, robust parameter density estimation, backend ranking, noise correlation, and resource estimation analysis. Instead of reporting only one best parameter vector, the framework identifies robust parameter regions, residual gaps to ideal behavior, backend specific failure modes, and calibration sensitive uncertainty. The result is a common benchmark for variational and non-variational workloads that measures how reliably each backend reaches useful task level behavior.
Quantum Machine Learning (QML) leverages quantum mechanical properties to enhance computational capabilities. With its emergence, there is a need to integrate QML models into machine learning pipelines for real-life applications such as image processing. While standalone programs exist to demonstrate the performance of QML models, a well-defined model workflow is noticeably absent. We present piQture, an open-source Python and Qiskit-based library that streamlines the implementation and training of QML models tailored for image processing. Its design and structure prioritize usability among users familiar with classical machine learning without prior QML experience.