Abstract:Generative models realized with machine learning techniques are powerful tools to infer complex and unknown data distributions from a finite number of training samples in order to produce new synthetic data. Diffusion models are an emerging framework that have recently overcome the performance of the generative adversarial networks in creating synthetic text and high-quality images. Here, we propose and discuss the quantum generalization of diffusion models, i.e., three quantum-noise-driven generative diffusion models that could be experimentally tested on real quantum systems. The idea is to harness unique quantum features, in particular the non-trivial interplay among coherence, entanglement and noise that the currently available noisy quantum processors do unavoidably suffer from, in order to overcome the main computational burdens of classical diffusion models during inference. Hence, we suggest to exploit quantum noise not as an issue to be detected and solved but instead as a very remarkably beneficial key ingredient to generate much more complex probability distributions that would be difficult or even impossible to express classically, and from which a quantum processor might sample more efficiently than a classical one. An example of numerical simulations for an hybrid classical-quantum generative diffusion model is also included. Therefore, our results are expected to pave the way for new quantum-inspired or quantum-based generative diffusion algorithms addressing more powerfully classical tasks as data generation/prediction with widespread real-world applications ranging from climate forecasting to neuroscience, from traffic flow analysis to financial forecasting.

A Quantum Machine Learning Algorithm Based on Generative Models

An efficient quantum algorithm for generative machine learning

Quantum Generative Models for Data Generation

Recent Advances for Quantum Neural Networks in Generative Learning

Power of Quantum Generative Learning

Quantum-Noise-Driven Generative Diffusion Models

Quantum generative adversarial network for generating discrete distribution

A quantum generative model for multi-dimensional time series using Hamiltonian learning

Quantum Deep Generative Prior with Programmable Quantum Circuits

Generative quantum machine learning via denoising diffusion probabilistic models

Learning and Inference on Generative Adversarial Quantum Circuits

Quantum Generative Model with Variable-Depth Circuit

Quantum Generative Diffusion Model: A Fully Quantum-Mechanical Model for Generating Quantum State Ensemble

A Generative Modeling Approach Using Quantum Gates

Quantum Machine Learning: Fad or Future?

Predicting Properties of Quantum Systems with Conditional Generative Models

Experimental Machine Learning of Quantum States

Power of data in quantum machine learning

LatentQGAN: A Hybrid QGAN with Classical Convolutional Autoencoder