aiqbench analysis
Stochastic Schrödinger Diffusion Models offer a significant advancement in quantum machine learning by enabling direct sampling within the complex projective manifold of pure-state representations. This framework bypasses classical input perturbations and utilizes an analytic teacher score to overcome transition density limitations inherent in previous Riemannian approaches, demonstrating superior performance on benchmarks up to 14 qubits. The method effectively improves representation-level diagnostics for downstream quantum kernel methods while maintaining fidelity across synthetic and physics-inspired ensembles.
Stochastic Schr\"odinger Diffusion Models for Pure-State Ensemble Generation
arXiv:2605.03573v3 Announce Type: replace-cross Abstract: Quantum machine learning increasingly relies on pure-state representations, motivating generative models that sample directly in quantum representation space rather than perturbing classical inputs and re-encoding. We introduce Stochastic Schr\"odinger Diffusion Models (SSDMs), a score-based generative framework that defines diffusion, scores, and reverse-time sampling intrinsically on the complex projective manifold $\mathbb{CP}^{d-1}$ under the Fubini--Study metric. SSDMs combine a Riemannian Ornstein--Uhlenbeck forward diffusion with a stochastic Schr\"odinger realization, and learn reverse-time dynamics driven by the Riemannian score. Our central technical contribution is a local-time learning objective that exploits the local Euclidean OU limit of intrinsic manifold diffusions in Fubini-Study normal coordinates to obtain an analytic teacher score, bypassing the intractable transition densities that limit existing Riemannian score-based models. Across synthetic, physics-inspired (TFIM, XXZ), and quantum feature-state benchmarks up to $14$ qubits, SSDMs match target pure-state ensembles by orders of magnitude on MMD and observable statistics over both ambient Euclidean and matched Riemannian score-based baselines, and improve representation-level diagnostics for downstream quantum kernel methods.
Abstract:Quantum machine learning increasingly relies on pure-state representations, motivating generative models that sample directly in quantum representation space rather than perturbing classical inputs and re-encoding. We introduce Stochastic Schrödinger Diffusion Models (SSDMs), a score-based generative framework that defines diffusion, scores, and reverse-time sampling intrinsically on the complex projective manifold $\mathbb{CP}^{d-1}$ under the Fubini--Study metric. SSDMs combine a Riemannian Ornstein--Uhlenbeck forward diffusion with a stochastic Schrödinger realization, and learn reverse-time dynamics driven by the Riemannian score. Our central technical contribution is a local-time learning objective that exploits the local Euclidean OU limit of intrinsic manifold diffusions in Fubini-Study normal coordinates to obtain an analytic teacher score, bypassing the intractable transition densities that limit existing Riemannian score-based models. Across synthetic, physics-inspired (TFIM, XXZ), and quantum feature-state benchmarks up to $14$ qubits, SSDMs match target pure-state ensembles by orders of magnitude on MMD and observable statistics over both ambient Euclidean and matched Riemannian score-based baselines, and improve representation-level diagnostics for downstream quantum kernel methods.
Machine Learning (stat.ML); Machine Learning (cs.LG)
arXiv:2605.03573 [stat.ML]
(or arXiv:2605.03573v3 [stat.ML] for this version)
Full text under the publisher's open licence · 175 words
Related coverage
Stochastic Schr\"odinger Equations for Quantum Reverse Diffusion
arXiv · Jun 4, 2026
Generating quantum ensembles via reverse-time quantum diffusions
arXiv · Jun 3, 2026
Reverse quantum state diffusion from differential geometry
arXiv · Oct 1, 2026
Generating quantum ensembles via reverse-time quantum diffusions
arXiv · Sep 21, 2026