Searcharxiv⌕ Search

arXiv subjects

Laia Domingo

Publications and source records attributed to Laia Domingo.

10 recordsLinked to original sources

Quantum-inspired orthogonal adapters for certifiably robust medical image classification

Medical image classifiers require not only high accuracy but also reliable predictions under small input perturbations. Certified robustness provides mathematical guarantees of prediction stability, but preserving these guarantees when adapting pretrained models to new tasks remains challenging. We propose a fine-tuning approach based on quantum-inspired orthogonal butterfly layers that preserves the Lipschitz bounds of pretrained certified networks. Inspired by quantum circuits of reconfigurable beam splitters, these layers preserve orthogonality by construction using only $\mathcal{O}(n\log n)$ trainable parameters. Their structured design enables efficient classical implementation and offers a path toward future quantum implementations with logarithmic circuit depth under suitable connectivity assumptions. Controlled experiments on MNIST and MedMNIST demonstrate improved certified robustness compared with conventional alternatives and, in the DermaMNIST transfer, greater resistance to the tested attacks than low-rank adaptation (LoRA). We then apply it to trichoscopic image classification for alopecia diagnosis and alopecia areata activity assessment. By adapting a certified ImageNet-pretrained network, butterfly fine-tuning achieves approximately twice the certified accuracy of conventional fine-tuning and LoRA, with comparable classification accuracy on the same backbone and data split and similar empirical resistance to the tested adversarial attacks. Larger robustness guarantees, including certificates covering certain image-processing transformations, can be obtained at the cost of reduced accuracy. These results highlight the potential of quantum-inspired butterfly layers for parameter-efficient adaptation of medical image classifiers while preserving the mathematical guarantees underlying certified robustness.

cs.CV↗

eigenkit: A Provider-Neutral Quantum Computing Library

Quantum software ecosystems provide increasingly capable tools for simulation and hardware execution, but differences in programming models, supported features, and backend capabilities can make it difficult to move and systematically compare the same quantum algorithm across execution environments. We introduce eigenkit, a provider-neutral quantum computing library that separates algorithm definition from backend execution through a common circuit representation and provider interface. The library currently supports 12 execution providers spanning simulators and quantum hardware, provides backend-independent gradient computation, and integrates with standard scientific-computing and machine-learning workflows through NumPy, PyTorch, and scikit-learn interfaces. We evaluate the practical relevance of this abstraction across both simulation and quantum hardware. Simulator performance varies substantially with circuit size, structure, batching, and application workload, showing that no single backend is consistently preferable. On quantum hardware, experiments on IBM Heron and IQM Emerald show that which processor is better suited to a given algorithm depends not only on hardware constraints, but also on how well the topology of the quantum algorithm matches the physical connectivity of the processor. These results demonstrate that backend portability is not only a software-design convenience, but can be important for both computational efficiency and quantum execution quality.

quant-ph↗

Classical training of conditional instantaneous quantum polynomial-time Born machines

Instantaneous quantum polynomial-time (IQP) Born machines are attractive quantum generative models, as certain IQP circuit families are conjectured to be hard to sample, yet they can be trained classically. Previously proposed conditional IQP Born machines make the parameters of the diagonal layer condition dependent. We introduce conditional IQP Born machines, in which the condition is encoded in the input state and a single IQP unitary is shared across conditions. Efficient sampling and amplitude-ratio evaluation for the reference state entering the diagonal layer suffice to obtain unbiased classical estimates of a conditional maximum mean discrepancy objective and its gradients. We show that input conditioning can represent conditional families that no diagonally conditioned IQP model with the same number of qubits can reproduce. For the sixteen-qubit Curie-Weiss spin model, a single conditional model captures the magnetization transition at held-out temperatures. Its mean total variation distance is six times lower than that of diagonal conditioning under the same protocol and comparable to that of classical generative models with up to seven times as many parameters. We also train conditional models classically with up to 2048 qubits, which reproduce the mean squared magnetization of the target away from the crossover. These results extend classically trained IQP Born machines to conditional generation.

quant-ph↗

Diagnosing quantum reservoirs at scale based on expressivity and coverage

Quantum reservoirs offer a hardware-friendly route to quantum machine learning, replacing trainable circuits with fixed random dynamics and a classical readout. Because the reservoir is not optimized, performance depends entirely on the choice of reservoir family, yet existing diagnostics demand resources that grow exponentially with system size. We introduce a scalable, hardware-agnostic framework built on two complementary quantities. The first is a task-independent order-statistics (ORS) expressivity score, which compares only the largest output probabilities of a reservoir ensemble against an analytical Haar baseline. It never reconstructs the full output distribution, is cost-independent of Hilbert-space dimension, and admits a closed-form depolarizing noise correction, making it directly usable on hardware. The second is the task-dependent effective rank $R_{\mathrm{eff}}$ of the feature matrix, which measures how much input-dependent information reaches the readout. We validate the ORS score against established complexity diagnostics and confirm it remains informative under simulated noise and on IBM quantum hardware. Across synthetic and real quantum extreme learning machine and quantum reservoir computing benchmarks, ORS captures the intrinsic expressivity hierarchy of reservoir families while $R_{\mathrm{eff}}$ determines when that expressivity becomes usable predictive information.

quant-ph↗

Superpixel-Based QUBO for Scalable Quantum-Enhanced Medical Image Segmentation

Quadratic unconstrained binary optimization (QUBO) has emerged as a powerful framework for medical computing problems. Binary decision variables naturally represent clinical choices, making QUBO formulations well-suited for quantum annealing hardware. However, a fundamental scalability challenge limits practical deployment: problem size grows rapidly with input dimensionality, creating computational bottlenecks that restrict applications to simplified scenarios. This paper addresses this challenge through hierarchical problem reduction, as demonstrated in medical image segmentation, where pixel-level QUBO formulations create over 65,000 variables for a 256x256 image, forcing existing approaches to downsample to 42x42 resolution and discard 97% of pixel information. A superpixel-based QUBO framework is proposed using simple linear iterative clustering (SLIC) to group pixels into perceptually meaningful regions, then formulate segmentation as QUBO over a region adjacency graph (RAG) combining min-cut and smoothness objectives. Validation on INbreast mammography breast cancer images demonstrates a 4.2% improvement in segmentation quality (mean IoU 0.76 vs 0.73) with 33 computational speedup (0.67s vs 21.97s) and a 97.3% reduction in problem size (1764 to 48 variables), all achieved while processing full-resolution images rather than downsampled versions. The reduced problem size also fits well within current quantum annealer connectivity limits, removing the embedding overhead that has historically blocked direct deployment of pixel-level QUBO segmentation on quantum hardware.

cs.CV↗

Quantum-enhanced optimization for patient stratification in clinical trials

Clinical trials are notorious for their high failure rates and steep costs, leading to wasted time and resources spend, prolonged development timelines, and delayed patient access to new therapies. A key contributor to these failures is biological uncertainty, which complicates trial design and weakens the ability to detect true treatment effects. In particular, inadequate patient stratification often results in covariate imbalances across treatment arms, masking treatment effects and reducing statistical power, even when therapies are effective for specific patient subpopulations. This work presents an optimization-based, quantum-enhanced approach to patient stratification that explicitly minimizes covariate imbalance across numerical and categorical variables, without altering protocol design or trial endpoints. Using real clinical trial data, we demonstrate that hybrid quantum-classical optimization methods achieve high-quality stratification while scaling efficiently to larger cohorts. In our benchmark study, the quantum-enhanced pipeline delivered over a 100x improvement in computational efficiency compared to classical approaches, enabling faster iteration and practical deployment at scale. This report shows how improved stratification can lead to decision-relevant gains, including up to a fivefold increase in statistical significance in treatment effect estimation, reducing treatment-effect dilution and increasing trial sensitivity. Together, these results show that optimization-driven stratification can strengthen clinical trial design, improve confidence in downstream decisions, and reduce the risk of costly late-stage failure.

quant-ph↗

Quantum-enhanced unsupervised image segmentation for medical images analysis

Breast cancer remains the leading cause of cancer-related mortality among women worldwide, necessitating the meticulous examination of mammograms by radiologists to characterize abnormal lesions. This manual process demands high accuracy and is often time-consuming, costly, and error-prone. Automated image segmentation using artificial intelligence offers a promising alternative to streamline this workflow. However, most existing methods are supervised, requiring large, expertly annotated datasets that are not always available, and they experience significant generalization issues. Thus, unsupervised learning models can be leveraged for image segmentation, but they come at a cost of reduced accuracy, or require extensive computational resourcess. In this paper, we propose the first end-to-end quantum-enhanced framework for unsupervised mammography medical images segmentation that balances between performance accuracy and computational requirements. We first introduce a quantum-inspired image representation that serves as an initial approximation of the segmentation mask. The segmentation task is then formulated as a QUBO problem, aiming to maximize the contrast between the background and the tumor region while ensuring a cohesive segmentation mask with minimal connected components. We conduct an extensive evaluation of quantum and quantum-inspired methods for image segmentation, demonstrating that quantum annealing and variational quantum circuits achieve performance comparable to classical optimization techniques. Notably, quantum annealing is shown to be an order of magnitude faster than the classical optimization method in our experiments. Our findings demonstrate that this framework achieves performance comparable to state-of-the-art supervised methods, including UNet-based architectures, offering a viable unsupervised alternative for breast cancer image segmentation.

eess.IV↗

Classical and quantum reservoir computing: development and applications in machine learning

Reservoir computing is a novel machine learning algorithm that uses a nonlinear dynamical system to efficiently learn complex temporal patterns from data. The objective of this thesis is to investigate the principles of reservoir computing and develop state-of-the-art variants capable of addressing diverse applications in machine learning. The research demonstrates the algorithm's robustness and adaptability across very different domains, including agricultural time series forecasting and the time propagation of quantum systems. The first contribution of this thesis consists in developing a reservoir computing-based methodology to predict future agricultural product prices, which is crucial for ensuring the sustainability of the food market. The next contribution of the thesis is devoted to solving the Schrödinger equation for complex quantum systems. A novel reservoir computing framework is proposed to efficiently propagate quantum wavefunctions in time, enabling the computation of all eigenstates of a quantum system within a specific energy range. This approach is used to study prominent systems in the field of quantum chemistry and quantum chaos. The last contribution of this thesis focuses on optimizing algorithm designs for quantum reservoir computing. The results demonstrate that families of quantum circuits with higher complexity, according to the majorization criterion, yield superior performance in quantum machine learning. Moreover, the impact of quantum noise on the algorithm performance is evaluated, revealing that the amplitude damping noise can actually be beneficial for the performance of quantum reservoir computing, while the depolarizing and phase damping noise should be prioritized for correction. Furthermore, the optimal design of quantum reservoirs is employed to construct a hybrid quantum-classical neural network that tackles a fundamental problem in drug design.

quant-ph↗

Quantum reservoir complexity by Krylov evolution approach

Quantum reservoir computing algorithms recently emerged as a standout approach in the development of successful methods for the NISQ era, because of its superb performance and compatibility with current quantum devices. By harnessing the properties and dynamics of a quantum system, quantum reservoir computing effectively uncovers hidden patterns in data. However, the design of the quantum reservoir is crucial to this end, in order to ensure an optimal performance of the algorithm. In this work, we introduce a precise quantitative method, with strong physical foundations based on the Krylov evolution, to assess the wanted good performance in machine learning tasks. Our results show that the Krylov approach to complexity strongly correlates with quantum reservoir performance, making it a powerful tool in the quest for optimally designed quantum reservoirs, which will pave the road to the implementation of successful quantum machine learning methods.

quant-ph↗

Deep learning methods for the computation of vibrational wavefunctions

In this paper we design and use two Deep Learning models to generate the ground and excited wavefunctions of different Hamiltonians suitable for the study the vibrations of molecular systems. The generated neural networks are trained with Hamiltonians that have analytical solutions, and ask the network to generalize these solutions to more complex Hamiltonian functions. This approach allows to reproduce the excited vibrational wavefunctions of different molecular potentials. All methodologies used here are data-driven, therefore they do not assume any information about the underlying physical model of the system. This makes this approach versatile, and can be used in the study of multiple systems in quantum chemistry.

physics.chem-ph↗