Existing theory of variational quantum learning characterises circuits through three complementary perspectives: representation, optimization, and statistical capacity, but does not ask whether a chosen data encoding is compatible with the learning task at hand. We develop a task-centric theory of learnability to answer this question. Building on the Fourier characterization of parameterized quantum circuits, we introduce spectral task-circuit alignment, a quantitative measure of the overlap between the hypothesis space generated by a data encoding and the spectral structure of a target function. We prove that this quantity determines an irreducible approximation floor that cannot be reduced by deeper ansätze, additional trainable parameters, improved optimization, or more training data. Combining this result with the Effective Dimension framework gives a two-stage characterization of learnability that separates approximation barriers from statistical barriers. The resulting theory further leads to a principled criterion for encoding design, showing that alignment is maximized by encodings whose accessible frequency sets capture the dominant Fourier components of the target. We illustrate the framework through an analytical study of the feature-map construction of Abbas et al. on the checkerboard classification target, demonstrating that the empirical advantage of the hard feature map follows directly from its superior task-circuit alignment. More broadly, this work introduces task compatibility as a fourth theoretical component of quantum machine learning, complementing expressibility, trainability, and statistical capacity. The resulting framework provides a unified foundation for analysing approximation, learnability, and encoding design before any circuit is trained.
Existing theory of variational quantum learning characterises circuits through three complementary perspectives: representation, optimization, and statistical capacity, but does not ask whether a chosen data encoding is compatible with the learning task at hand. We develop a task-centric theory of learnability to answer this question. Building on the Fourier characterization of parameterized quantum circuits, we introduce spectral task-circuit alignment, a quantitative measure of the overlap between the hypothesis space generated by a data encoding and the spectral structure of a target function. We prove that this quantity determines an irreducible approximation floor that cannot be reduced by deeper ansätze, additional trainable parameters, improved optimization, or more training data. Combining this result with the Effective Dimension framework gives a two-stage characterization of learnability that separates approximation barriers from statistical barriers. The resulting theory further leads to a principled criterion for encoding design, showing that alignment is maximized by encodings whose accessible frequency sets capture the dominant Fourier components of the target. We illustrate the framework through an analytical study of the feature-map construction of Abbas et al. on the checkerboard classification target, demonstrating that the empirical advantage of the hard feature map follows directly from its superior task-circuit alignment. More broadly, this work introduces task compatibility as a fourth theoretical component of quantum machine learning, complementing expressibility, trainability, and statistical capacity. The resulting framework provides a unified foundation for analysing approximation, learnability, and encoding design before any circuit is trained.
This chapter develops a learning-theoretical foundation for quantum machine learning (QML) by treating quantum models as statistical learning systems rather than merely as computational circuits or algorithms. The discussion begins by motivating the need for learning theory in QML, emphasizing that nominal expressivity in exponentially large Hilbert spaces does not automatically imply learnability or generalization. Classical concepts such as VC dimension and Rademacher complexity are then adapted to quantum settings to analyze the effective capacity of parameterized quantum circuits and quantum kernels. These tools show that accessible quantum hypothesis classes are far more constrained than naive Hilbert-space arguments suggest and that increased expressivity can easily lead to overfitting when not balanced by appropriate inductive bias and regularization. This chapter further examines generalization bounds for variational quantum circuits, highlighting the additional roles of measurement-induced stochasticity, finite-shot variance, and hardware noise. Overfitting, implicit and explicit regularization, and the possibility of double-descent-like behavior in quantum models are analyzed from a unified statistical perspective. A final synthesis shows that expressivity, trainability, and robustness are fundamentally coupled in QML. Hands-on examples in Qiskit and PennyLane illustrate kernel complexity, generalization gaps, and depth-dependent trainability. Overall, this chapter provides a principled framework for evaluating when quantum models can learn reliably from finite, noisy data.
Parameterized quantum circuits are widely investigated in quantum machine learning, yet how their architecture can provide task-aligned inductive biases remains poorly understood. We study data reuploading quantum circuits with two distinct inputs: a signal that varies between samples and a fixed, task-defined spectral query that specifies how the signal should be probed. We show that encoding the query can change the circuit's hypothesis class over the signal by modulating how its existing signal frequencies are combined, without introducing new sample-dependent information. This mechanism matches tasks built on nonuniform Fourier transforms, whose evaluation frequencies specify the query and whose sampling positions guide its encoding. We examine this connection in a classification task inspired by synthetic aperture radar, where acquisition geometry determines a query for classifying unfocused spectral patches before image formation. Across small-scale spectral and radar-like tasks, aligned queries improve learning from limited data over parameter-matched circuits with mismatched or absent queries. These results provide a route from known spectral task structure to the design and validation of inductive biases in parameterized quantum circuits.
In this work, we investigate the phenomenon of spectral bias in quantum machine learning, where, in classical settings, models tend to fit low-frequency components of a target function earlier during training than high-frequency ones, demonstrating a frequency-dependent rate of convergence. We study this effect specifically in parameterised quantum circuits (PQCs). Leveraging the established formulation of PQCs as Fourier series, we prove that spectral bias in this setting arises from the ``redundancy'' of the Fourier coefficients, which denotes the number of terms in the analytical form of the model contributing to the same frequency component. The choice of data encoding scheme dictates the degree of redundancy for a Fourier coefficient. We find that the magnitude of the Fourier coefficients' gradients during training strongly correlates with the coefficients' redundancy. We then further demonstrate this empirically with three different encoding schemes. Additionally, we demonstrate that PQCs with greater redundancy exhibit increased robustness to random perturbations in their parameters at the corresponding frequencies. We investigate how design choices affect the ability of PQCs to learn Fourier sums, focusing on parameter initialization scale and entanglement structure, finding large initializations and low-entanglement schemes tend to slow convergence.
Parametrised quantum circuits learn by adjusting gate parameters, while their design shapes the functions they can represent and how readily they learn them. We introduce the circuit harmonic matrix, a fixed matrix organising Fourier expansions over inputs and parameters. It makes the effects of encoding, gates, initial state and observable explicit in coefficient variance, covariance and the quantum neural tangent kernel, linking variation across parameter space to local sensitivity. The construction also constrains representable functions and lower-bounds fitting error. For fixed Clifford gates and independently parametrised Pauli or controlled rotations, we develop a two-copy propagation method that computes covariance and the averaged tangent kernel analytically for uniform parameters, without constructing the full matrix or sampling parameters. Across 400 configurations spanning eight circuit families, we compare these quantities with actual learning. Larger coefficient variance is strongly associated with faster learning and lower error for the corresponding target frequency, revealing spectral bias towards lower frequencies. Increasing depth and qubit count generally suppresses total variance, yet deeper circuits usually learn faster and achieve lower error, while adding qubits slows early learning and has a frequency-dependent effect on final error. Matched targets show no robust overall advantage from covariance alignment. For one circuit setting, initial tangent kernels closely forecast the modest loss reductions reached in small-step gradient descent. These results connect circuit choices to representational constraints and learning performance, providing a basis for selecting designs suited to particular learning tasks.
This chapter develops variational quantum models as the dominant framework for near-term quantum machine learning (QML), emphasizing their role as trainable quantum circuits embedded within hybrid quantum-classical optimization loops. Building on earlier chapters on encoding and quantum feature geometry, it formulates parameterized quantum circuits (PQCs) as families of unitary transformations acting on reference states, with model outputs obtained through expectation values of observables. This measurement-based structure makes variational learning inherently stochastic and closely ties model behavior to circuit architecture, observable choice, and optimization strategy. This chapter examines variational quantum circuits (VQCs) as learning models, introduces quantum neural networks (QNNs) as structured variational architectures, and analyzes layered, data-reuploading, recurrent, and hybrid classical-quantum designs. It highlights the central tension between expressivity and trainability, showing how universality and approximation power are counterbalanced by optimization barriers such as barren plateaus. The discussion further develops ansatz design strategies, supervised, fidelity-based, and energy-based losses, regularization under noise, and practical methods for gradient estimation and optimization, including the parameter-shift rule, Simultaneous Perturbation Stochastic Approximation (SPSA), and layer-wise training. Together, these ideas establish variational quantum models as a flexible yet hardware-constrained paradigm whose practical success depends on careful coordination among architecture, measurement, and optimization.
Variational quantum algorithms (VQAs) have shown advantages over classical approaches in solving many problems, yet their quantum neural networks (QNNs) models still perform poorly in complex tasks like multi-class classification and high-frequency time-series prediction, raising questions about their feasibility. While existing studies have characterized their performance from several perspectives, but research on their learning performance and visualization remains scarce. This work develops the universal approximation theory of variational quantum algorithms for multivariate function fitting, supported by a verification in visual training analysis on two-input binary datasets. We first visualize how a single-qubit VQA with single-encoding process four types of binary data, revealing inherent limitations, and then formulate the measurement expectation of VQA as a partial Fourier series, establishing its universal approximation capacity and showing that the Fourier spectrum is determined by the data encoding strategy: multiple encoding layers broaden the spectrum and enhance learning performance. Guided by this theory, we adjust the circuit structure and achieve nearly 100% accuracy across all four binary classification tasks.