arXiv subjects

Danqi Zhuang

Publications and source records attributed to Danqi Zhuang.

3 recordsLinked to original sources

arXiv2026

On the computational cost of Stochastic Gradient Langevin Dynamics

Stochastic Gradient Langevin Dynamics (SGLD) reduces the cost of Langevin-based sampling by replacing full-dataset drift evaluations with mini-batch approximations, but the resulting subsampling error may offset this computational saving. We study this trade-off for stochastic differential equations with finite-sum drifts and compare the computational cost of SGLD with that of the Euler-Maruyama (EM) method. For a prescribed mean-square accuracy $\varepsilon^2$, we derive complexity estimates that explicitly track the dependence on the dataset size $m$, mini-batch size $s$, and accuracy parameter $\varepsilon$. The resulting comparison reveals distinct parameter regimes in which either method is preferable. In particular, EM can have lower leading-order cost only in a small-data, aggressive-subsampling regime, whereas SGLD is favoured over most of the remaining parameter space. In the practically relevant regime $s \ll m$, the transition between the two methods occurs at the scale $m \asymp \varepsilon^{-1}$. We complement the theoretical analysis with numerical experiments based on a Gaussian Bayesian inference model, which examine the predicted cost regimes together with the underlying discretisation error and variance estimates.

math.NA↗

arXiv2026

PTL-Diffusion: Manifold-Aware Diffusion with Periodic Terminal Laws

Standard diffusion models typically use a single time-homogeneous Gaussian terminal distribution as the reference law for generation. While this choice is analytically convenient and empirically powerful, it provides little explicit structure for data concentrated near low-dimensional manifolds, where different regions of the data distribution may correspond to distinct local geometric or semantic factors. As a result, the reverse model must recover manifold-level structure almost entirely from an unstructured terminal reference distribution. We propose PTL-Diffusion, a proof-of-concept diffusion framework whose forward noising process converges to a nonconstant periodic family of Gaussian terminal laws rather than to a single invariant law. Unlike a phase-conditioned DDPM, where phase information only enters the denoising network while the forward process remains unchanged, PTL-Diffusion embeds phase structure directly into the forward noising dynamics. The proposed construction remains close to standard denoising diffusion models: for a periodically forced Ornstein--Uhlenbeck-type forward process, we derive closed-form forward marginals, the limiting periodic Gaussian terminal family, and explicit Gaussian reverse posteriors, enabling standard noise-prediction training. We also introduce an invariant-average regularization term coupling the phase-conditioned reverse dynamics through the averaged periodic reference law. Experiments on torus and cylinder point-cloud benchmarks and the Olivetti face dataset show that PTL-Diffusion improves manifold-level distributional matching over matched DDPM baselines, reducing phase-conditioned errors, feature-space covariance errors, and nearest-neighbour manifold distances. These results suggest structured terminal reference laws as a promising direction, while motivating more expressive phase constructions and larger-scale evaluations.

cs.CV↗

arXiv2025

Randomised Euler-Maruyama Method for SDEs with Hölder Continuous Drift Coefficient Driven by $α$-stable Lévy Process

In this paper, we examine the performance of randomised Euler-Maruyama (EM) method for additive time-inhomogeneous SDEs with an irregular drift driven by symmetric $α$-table process, $α\in (1,2)$. In particular, the drift is assumed to be $β$-Hölder continuous in time and bounded $η$-Hölder continuous in space with $β,η\in (0,1]$. The strong order of convergence of the randomised EM in $L^p$-norm is shown to be $1/2+(β\wedge (η/α)\wedge(1/2))-\varepsilon$ for an arbitrary $\varepsilon\in (0,1/2)$, higher than the one of standard EM, which cannot exceed $β$. The result for the case of $α\in (1,2)$ extends the almost optimal order of convergence of randomised EM obtained in (arXiv:2501.15527) for SDEs driven by Gaussian noise ($α=2$), and coincides with the performance of EM method in simulating time-homogenous SDEs driven by $α$-stable process considered in (arXiv:2208.10052). Various experiments are presented to validate the theoretical performance.

math.PR↗