Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

arXiv:2201.09134 · Machine Learning Science and Technology

Data-centric machine learning in quantum information science

Sanjaya Lohani, Joseph M. Lukens, Ryan T. Glasser, Thomas A. Searles, Brian T. Kirby

Published Sep 7, 2022• 16 citations
arXivHTMLPDF

Abstract

Abstract We propose a series of data-centric heuristics for improving the performance of machine learning systems when applied to problems in quantum information science. In particular, we consider how systematic engineering of training sets can significantly enhance the accuracy of pre-trained neural networks used for quantum state reconstruction without altering the underlying architecture. We find that it is not always optimal to engineer training sets to exactly match the expected distribution of a target scenario, and instead, performance can be further improved by biasing the training set to be slightly more mixed than the target. This is due to the heterogeneity in the number of free variables required to describe states of different purity, and as a result, overall accuracy of the network improves when training sets of a fixed size focus on states with the least constrained free variables. For further clarity, we also include a ‘toy model’ demonstration of how spurious correlations can inadvertently enter synthetic data sets used for training, how the performance of systems trained with these correlations can degrade dramatically, and how the inclusion of even relatively few counterexamples can effectively remedy such problems.

DOI
10.1088/2632-2153/ac9036

Related papers

Learning complexity gradually in quantum machine learning models

Erik Recio-Armengol, Franz J. Schreiber, Jens Eisert +1 more

Quantum machine learning is an emergent field that continues to draw significant interest for its potential to offer improvements over classical algorithms in certain areas. However, training quantum models remains a challenging task, largely because of the difficulty in establishing an effective inductive bias when solving high-dimensional problems. In this work, we propose a training framework that prioritizes informative data points over the entire training set. This approach draws inspiration from classical techniques such as and to introduce an additional inductive bias through the training data itself. By selectively focusing on informative samples, we aim to steer the optimization process toward more favorable regions of the parameter space. This data-centric approach complements existing strategies such as warm-start initialization methods, providing an additional pathway to address performance challenges in quantum machine learning. We provide theoretical insights into the benefits of prioritizing informative data for quantum models, and we validate our methodology with numerical experiments on selected recognition tasks of quantum phases of matter. Our findings indicate that this strategy could be a valuable approach for improving the performance of quantum machine learning models.

arXiv:2411.11954•Physical Review Research•May 19, 2026•4 citations

Data-Centric Quantum Machine Learning: Data-to-Models Continuum

Marcelo Rita Pias, Everson da Silva Flores, Bruna de Vargas Guterres +2 more

Preprint Quantum machine learning (QML) is a machine learning approach that uses quantum computation to improve learning performance. Research in this area has primarily followed a model-centric approach. However, near-term quantum hardware shows that data preparation also plays an important role, and can influence the quality and scalability of the learning process. Existing QML research focuses on circuits, kernels, and variational models; however, the data requirements that shape these methods remain less explored. This topical review addresses two questions. First, what forms of data processing are required across representative QML methods? Second, how can a data-centric approach improve the accuracy, scalability and resource use in QML models? A taxonomy was introduced to classify QML architectures according to their requirements for dimensionality reduction, encoding, dataset balance and compression. Based on this taxonomy, a data-centric QML (dc-qml) framework is proposed. It integrates various subprocesses into a coherent data flow, such as data ingestion, abstraction, tokenization, quantum encoding, and synthetic data generation. Results from recent benchmarks and case studies show that improvements in near-term QML optimizations often stem from data-focused practices rather than model refinement alone.

IEEE Access•Jan 1, 2026•0 citations

Machine learning meeting the Clauser-Horne-Shimony-Holt scenario

Gabriel Pereira Alves, N. Gigena, Jędrzej Kaniewski

In this work we perform a comprehensive study of the machine learning (ML) methods for the purpose of characterizing the quantum set of correlations. As our main focus is on assessing the usefulness and effectiveness of the ML approach, we focus exclusively on the Clauser-Horne-Shimony-Holt scenario, both the four-dimensional variant, for which an analytical solution is known, and the eight-dimensional variant, for which no analytical solution is known but numerical approaches are relatively well understood. We consider a wide selection of approaches, ranging from simple data science models to dense neural networks. The two classes of models that perform well are support vector machines and dense neural networks, and they are the main focus of this work. We conclude that while it is relatively easy to achieve good performance on average, it is hard to train a model that performs well on the hard cases, i.e., points in the vicinity of the boundary of the quantum set, which are precisely the cases that are interesting from the academic point of view. In order to improve performance on hard cases we must, especially for the eight-dimensional problem, resort to a tailored choice of training data, which means that we are implicitly feeding our intuition and biases into the model. We feel that this is an important and often overlooked aspect of applying ML models to academic problems, where data generation or data selection is performed according to some implicit subjective criteria. In this way, it is possible to unconsciously steer our model so that it exhibits features that we are interested in seeing. Hence, special care must be taken while determining whether ML methods can be considered objective and unbiased in the context of academic problems.

arXiv:2407.14396•Physical Review A•Mar 17, 2025•1 citation

Quantum-Enhanced Machine Learning

Vedran Dunjko, Jacob M. Taylor, Hans J. Briegel

The emerging field of quantum machine learning has the potential to substantially aid in the problems and scope of artificial intelligence. This is only enhanced by recent successes in the field of classical machine learning. In this work we propose an approach for the systematic treatment of machine learning, from the perspective of quantum information. Our approach is general and covers all three main branches of machine learning: supervised, unsupervised, and reinforcement learning. While quantum improvements in supervised and unsupervised learning have been reported, reinforcement learning has received much less attention. Within our approach, we tackle the problem of quantum enhancements in reinforcement learning as well, and propose a systematic scheme for providing improvements. As an example, we show that quadratic improvements in learning efficiency, and exponential improvements in performance over limited time periods, can be obtained for a broad class of learning problems.

arXiv:1610.08251•Physical Review Letters•Sep 20, 2016•491 citations

Quantum Information Framework for Neural Network Generalization: A Comprehensive Experimental Analysis

Lam, Chi Hin

The remarkable generalization capabilities of overparameterized neural networks remain one of the most profound mysteries in modern machine learning. Despite having sufficient capacity to simply memorize training data, these networks consistently achieve superior performance on unseen test examples. Traditional analyses focusing on weight distributions, loss landscapes, and optimization dynamics have provided partial insights but fail to capture the complex correlational structure that emerges between neurons during learning. This paper introduces a comprehensive quantum information framework for analyzing neural network representations, treating hidden layer activations as quantum states and characterizing them through density matrices, von Neumann entropy, purity, and related quantum information metrics. Through systematic experiments on synthetic classification tasks and algorithmic reasoning problems spanning over 3000 cumulative training epochs and 5000 Wang-Landau sampling iterations, we demonstrate that these quantum metrics reveal structure completely invisible to classical analysis. Our results show that stochastic gradient descent systematically reduces von Neumann entropy while increasing purity as networks learn, transitioning from highly mixed quantum states with entropy approximately 1.42 and purity approximately 0.29 to purer configurations with entropy approximately 1.19 and purity approximately 0.37 that achieve superior test accuracy of 99.5%. Using Wang-Landau sampling to explore the equilibrium entropy landscape, we discover that neural networks naturally favor high-entropy states with entropy approximately 1.42 and moderate accuracy of 62.0%, and that optimization actively purifies the quantum state to achieve task performance. On modular arithmetic tasks exhibiting the grokking phenomenon, we observe that von Neumann entropy increases dramatically from 1.55 to 3.29, representing a 112% increase, during an extended 2000-epoch memorization phase before any improvement in generalization, while purity decreases correspondingly from 0.444 to 0.042, a 90.5% decrease, approaching the maximally mixed limit of 0.01. This entropy accumulation phase appears necessary for eventual generalization, with the network building quantum correlations at a rate of approximately 0.00097 nats per epoch during the plateau. Weight entropy, in contrast, shows minimal change of only 11% increase over the same period. Correlation analysis reveals that von Neumann entropy and purity are almost perfectly anticorrelated with a correlation coefficient of 1-0.99, while weight entropy shows weak correlation with quantum metrics with absolute correlation coefficients less than 0.3. The effective rank of the density matrix stabilizes near 2.7 to 3.0 for the 6-neuron classifier, indicating that the network effectively uses approximately three independent activation modes. These findings establish quantum information metrics as powerful tools for understanding neural network learning dynamics, provide the first quantitative evidence for the entropy accumulation interpretation of grokking, and suggest that neural networks undergo a quantum-inspired phase transition during learning. We discuss implications for generalization theory, optimization algorithm design, and the statistical physics of deep learning, and outline a comprehensive research agenda for extending these methods to larger architectures and more complex tasks.

•Mar 18, 2026•0 citations

dc-qml: data-centric quantum machine learning

Everson da Silva Flores, Kauã Dalla Riva Cucco Barbosa, Bruna de Vargas Guterres +2 more

Quantum machine learning research (QML) has largely followed a model-centric perspective that focuses on circuit design and algorithm refinement. However, near-term quantum hardware constraints place equal importance on how data are prepared and encoded before quantum processing. This short paper presents preliminary work on a data-centric quantum machine learning methodology termed dc-qml. The approach structures data ingestion, tokenization, encoding and synthetic data generation within a unified data-to-model workflow. The goal is to explore how systematic data preparation can improve the performance of QML systems. This work in progress is intended to generate discussion at the first SBCCQ workshop.

Simpósio Brasileiro de Computação e Comunicação Quânticas (SBCCQ)•Jul 19, 2026•0 citations
Abstract We propose a series of data-centric heuristics for improving the performance of machine learning systems when applied to problems in quantum information science. In particular, we consider how systematic engineering of training sets can significantly enhance the accuracy of pre-trained neural networks used for quantum state reconstruction without altering the underlying architecture. We find that it is not always optimal to engineer training sets to exactly match the expected distribution of a target scenario, and instead, performance can be further improved by biasing the training set to be slightly more mixed than the target. This is due to the heterogeneity in the number of free variables required to describe states of different purity, and as a result, overall accuracy of the network improves when training sets of a fixed size focus on states with the least constrained free variables. For further clarity, we also include a ‘toy model’ demonstration of how spurious correlations can inadvertently enter synthetic data sets used for training, how the performance of systems trained with these correlations can degrade dramatically, and how the inclusion of even relatively few counterexamples can effectively remedy such problems.
Quantum machine learning is an emergent field that continues to draw significant interest for its potential to offer improvements over classical algorithms in certain areas. However, training quantum models remains a challenging task, largely because of the difficulty in establishing an effective inductive bias when solving high-dimensional problems. In this work, we propose a training framework that prioritizes informative data points over the entire training set. This approach draws inspiration from classical techniques such as and to introduce an additional inductive bias through the training data itself. By selectively focusing on informative samples, we aim to steer the optimization process toward more favorable regions of the parameter space. This data-centric approach complements existing strategies such as warm-start initialization methods, providing an additional pathway to address performance challenges in quantum machine learning. We provide theoretical insights into the benefits of prioritizing informative data for quantum models, and we validate our methodology with numerical experiments on selected recognition tasks of quantum phases of matter. Our findings indicate that this strategy could be a valuable approach for improving the performance of quantum machine learning models.
Preprint Quantum machine learning (QML) is a machine learning approach that uses quantum computation to improve learning performance. Research in this area has primarily followed a model-centric approach. However, near-term quantum hardware shows that data preparation also plays an important role, and can influence the quality and scalability of the learning process. Existing QML research focuses on circuits, kernels, and variational models; however, the data requirements that shape these methods remain less explored. This topical review addresses two questions. First, what forms of data processing are required across representative QML methods? Second, how can a data-centric approach improve the accuracy, scalability and resource use in QML models? A taxonomy was introduced to classify QML architectures according to their requirements for dimensionality reduction, encoding, dataset balance and compression. Based on this taxonomy, a data-centric QML (dc-qml) framework is proposed. It integrates various subprocesses into a coherent data flow, such as data ingestion, abstraction, tokenization, quantum encoding, and synthetic data generation. Results from recent benchmarks and case studies show that improvements in near-term QML optimizations often stem from data-focused practices rather than model refinement alone.
In this work we perform a comprehensive study of the machine learning (ML) methods for the purpose of characterizing the quantum set of correlations. As our main focus is on assessing the usefulness and effectiveness of the ML approach, we focus exclusively on the Clauser-Horne-Shimony-Holt scenario, both the four-dimensional variant, for which an analytical solution is known, and the eight-dimensional variant, for which no analytical solution is known but numerical approaches are relatively well understood. We consider a wide selection of approaches, ranging from simple data science models to dense neural networks. The two classes of models that perform well are support vector machines and dense neural networks, and they are the main focus of this work. We conclude that while it is relatively easy to achieve good performance on average, it is hard to train a model that performs well on the hard cases, i.e., points in the vicinity of the boundary of the quantum set, which are precisely the cases that are interesting from the academic point of view. In order to improve performance on hard cases we must, especially for the eight-dimensional problem, resort to a tailored choice of training data, which means that we are implicitly feeding our intuition and biases into the model. We feel that this is an important and often overlooked aspect of applying ML models to academic problems, where data generation or data selection is performed according to some implicit subjective criteria. In this way, it is possible to unconsciously steer our model so that it exhibits features that we are interested in seeing. Hence, special care must be taken while determining whether ML methods can be considered objective and unbiased in the context of academic problems.
The emerging field of quantum machine learning has the potential to substantially aid in the problems and scope of artificial intelligence. This is only enhanced by recent successes in the field of classical machine learning. In this work we propose an approach for the systematic treatment of machine learning, from the perspective of quantum information. Our approach is general and covers all three main branches of machine learning: supervised, unsupervised, and reinforcement learning. While quantum improvements in supervised and unsupervised learning have been reported, reinforcement learning has received much less attention. Within our approach, we tackle the problem of quantum enhancements in reinforcement learning as well, and propose a systematic scheme for providing improvements. As an example, we show that quadratic improvements in learning efficiency, and exponential improvements in performance over limited time periods, can be obtained for a broad class of learning problems.
The remarkable generalization capabilities of overparameterized neural networks remain one of the most profound mysteries in modern machine learning. Despite having sufficient capacity to simply memorize training data, these networks consistently achieve superior performance on unseen test examples. Traditional analyses focusing on weight distributions, loss landscapes, and optimization dynamics have provided partial insights but fail to capture the complex correlational structure that emerges between neurons during learning. This paper introduces a comprehensive quantum information framework for analyzing neural network representations, treating hidden layer activations as quantum states and characterizing them through density matrices, von Neumann entropy, purity, and related quantum information metrics. Through systematic experiments on synthetic classification tasks and algorithmic reasoning problems spanning over 3000 cumulative training epochs and 5000 Wang-Landau sampling iterations, we demonstrate that these quantum metrics reveal structure completely invisible to classical analysis. Our results show that stochastic gradient descent systematically reduces von Neumann entropy while increasing purity as networks learn, transitioning from highly mixed quantum states with entropy approximately 1.42 and purity approximately 0.29 to purer configurations with entropy approximately 1.19 and purity approximately 0.37 that achieve superior test accuracy of 99.5%. Using Wang-Landau sampling to explore the equilibrium entropy landscape, we discover that neural networks naturally favor high-entropy states with entropy approximately 1.42 and moderate accuracy of 62.0%, and that optimization actively purifies the quantum state to achieve task performance. On modular arithmetic tasks exhibiting the grokking phenomenon, we observe that von Neumann entropy increases dramatically from 1.55 to 3.29, representing a 112% increase, during an extended 2000-epoch memorization phase before any improvement in generalization, while purity decreases correspondingly from 0.444 to 0.042, a 90.5% decrease, approaching the maximally mixed limit of 0.01. This entropy accumulation phase appears necessary for eventual generalization, with the network building quantum correlations at a rate of approximately 0.00097 nats per epoch during the plateau. Weight entropy, in contrast, shows minimal change of only 11% increase over the same period. Correlation analysis reveals that von Neumann entropy and purity are almost perfectly anticorrelated with a correlation coefficient of 1-0.99, while weight entropy shows weak correlation with quantum metrics with absolute correlation coefficients less than 0.3. The effective rank of the density matrix stabilizes near 2.7 to 3.0 for the 6-neuron classifier, indicating that the network effectively uses approximately three independent activation modes. These findings establish quantum information metrics as powerful tools for understanding neural network learning dynamics, provide the first quantitative evidence for the entropy accumulation interpretation of grokking, and suggest that neural networks undergo a quantum-inspired phase transition during learning. We discuss implications for generalization theory, optimization algorithm design, and the statistical physics of deep learning, and outline a comprehensive research agenda for extending these methods to larger architectures and more complex tasks.
Quantum machine learning research (QML) has largely followed a model-centric perspective that focuses on circuit design and algorithm refinement. However, near-term quantum hardware constraints place equal importance on how data are prepared and encoded before quantum processing. This short paper presents preliminary work on a data-centric quantum machine learning methodology termed dc-qml. The approach structures data ingestion, tokenization, encoding and synthetic data generation within a unified data-to-model workflow. The goal is to explore how systematic data preparation can improve the performance of QML systems. This work in progress is intended to generate discussion at the first SBCCQ workshop.