"Train classical, deploy quantum" requires rethinking generalization

• URL: http://arxiv.org/abs/2608.31117v1
• Date: Mon, 31 Aug 2026 17:26:27 GMT
• Title: "Train classical, deploy quantum" requires rethinking generalization
• Abstract summary: We benchmark a broad set of quantum and classical generative models by direct sampling.<n>We show that models trained with a moment-matching loss generally show worse generalization than the likelihood-trained models.
• Score: 42.16343861129583
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Generative models have become central across science and industry, from image and text synthesis to the design of molecules and materials. Quantum generative models are considered one of the most promising applications for quantum computers, since a quantum circuit naturally produces samples from the distribution it encodes, and for suitable circuits that distribution is believed to be hard for any classical computer to reproduce. A leading strategy trains these models on a classical computer and reserves the quantum device for generating samples at deployment. This is possible when the training loss can be evaluated on a classical computer. A prime example is the maximum mean discrepancy (MMD$^2$), a moment-matching loss that compares the model and the data through their Pauli-$Z$ correlations. Research so far has asked whether such models can be trained and whether their sampling is hard; whether minimizing such an objective yields a model that generalizes, rather than one that merely reproduces the training statistics, remains poorly understood. We benchmark a broad set of quantum and classical generative models by direct sampling and show that models trained with a moment-matching loss generally show worse generalization than the likelihood-trained models. We show this on two application-inspired datasets: first a cardinality-constrained dataset at up to $30$ qubits and second a dataset of genomic single-nucleotide variants, whose valid set is the observed data. These results indicate that a converged moment-matching loss is not a reliable measure of generalization, and that train-classical, deploy-quantum workflows will need approaches that target generalization directly, leaving open whether better training objectives suffice or whether the model architectures themselves must change.

Related papers

• Quantum Incremental Learning with Mixed State Prototypes [53.319426778720434]
  We introduce a novel quantum incremental learning framework based on trainable mixed-state prototypes.<n>Our model achieves high-dimensional feature concentration using a minimal number of qubits.
  arXiv  Detail & Related papers  (2026-08-11T04:22:42Z)
• Quantum Fourier Generative Models Trainable at Large Scale [13.839852906168787]
  We propose an algorithmic framework for building and training quantum generative models.<n>We develop a distinct training strategy where training is enabled at large scale by log-likelihood loss.<n>We demonstrate the efficiency of the suggested framework by validating loss estimation at the scale of over 1000 qubits.
  arXiv  Detail & Related papers  (2026-06-26T18:00:00Z)
• Multi-Objective Learning for Diffusion Models: A Statistical Theory under Semi-Supervised Learning [83.65494573724904]
  We develop a principled MOL framework for diffusion models with limited data.<n>We propose a two-stage training procedure that first fits lightweight specialist models from limited paired data.<n>We establish bounds generalization showing that the required number of paired samples only depends on the complexity of the specialist model classes.
  arXiv  Detail & Related papers  (2026-05-24T18:19:40Z)
• Encoding Numerical Data for Generative Quantum Machine Learning [0.820994582033002]
  Real-world data is often numeric, requiring models to translate between binary and continuous representations.<n>We show that it requires the models to learn correlations that are solely an artifact of the way the data is encoded, and not related to the data itself.<n>We propose a strategy based on Gray-codes that can be implemented with essentially no overhead, conserves structures in the data, and avoids artificial correlations in situations in which the standard approach creates them.
  arXiv  Detail & Related papers  (2026-03-24T16:39:58Z)
• Limits of quantum generative models with classical sampling hardness [2.321580694317368]
  We study quantum generative models from the perspective of output distributions.<n>We find that models that anticoncentrate are not trainable on average, including those exhibiting quantum advantage.<n>We conclude that quantum advantage can still be found in generative models, although its source must be distinct from anticoncentration.
  arXiv  Detail & Related papers  (2025-12-31T11:40:50Z)
• Photonic quantum generative adversarial networks for classical data [0.0]
  In generative learning, models are trained to produce new samples that follow the distribution of the target data.<n>We present a quantum GAN based on linear optical circuits and Fock-space encoding.<n>We demonstrate that the model can learn to generate images by training the model end-to-end experimentally on a single-photon quantum processor.
  arXiv  Detail & Related papers  (2024-05-09T18:00:10Z)
• Predictive Models from Quantum Computer Benchmarks [0.0]
  holistic benchmarks for quantum computers are essential for testing and summarizing the performance of quantum hardware. We introduce a general framework for building predictive models from benchmarking data using capability models. Our case studies use data from cloud-accessible quantum computers and simulations of noisy quantum computers.
  arXiv  Detail & Related papers  (2023-05-15T17:00:23Z)
• A Framework for Demonstrating Practical Quantum Advantage: Racing Quantum against Classical Generative Models [62.997667081978825]
  We build over a proposed framework for evaluating the generalization performance of generative models. We establish the first comparative race towards practical quantum advantage (PQA) between classical and quantum generative models. Our results suggest that QCBMs are more efficient in the data-limited regime than the other state-of-the-art classical generative models.
  arXiv  Detail & Related papers  (2023-03-27T22:48:28Z)
• Generalization Metrics for Practical Quantum Advantage in Generative Models [68.8204255655161]
  Generative modeling is a widely accepted natural use case for quantum computers. We construct a simple and unambiguous approach to probe practical quantum advantage for generative modeling by measuring the algorithm's generalization performance. Our simulation results show that our quantum-inspired models have up to a $68 times$ enhancement in generating unseen unique and valid samples.
  arXiv  Detail & Related papers  (2022-01-21T16:35:35Z)
• Goal-directed Generation of Discrete Structures with Conditional Generative Models [85.51463588099556]
  We introduce a novel approach to directly optimize a reinforcement learning objective, maximizing an expected reward. We test our methodology on two tasks: generating molecules with user-defined properties and identifying short python expressions which evaluate to a given target value.
  arXiv  Detail & Related papers  (2020-10-05T20:03:13Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.