My research interests lie in machine learning algorithms and theory, with a particular focus on MCMC sampling algorithms, diffusion inference and modeling, optimization, and mean-field analysis.
selected publications
COLT
Almost linear convergence under minimal score assumptions: Quantized transition diffusion
Continuous diffusion models have demonstrated remarkable generative performance across diverse domains but are often constrained by the computational cost of simulating reverse Ornstein–Uhlenbeck processes via SDE/ODE solvers. Existing theoretical results typically establish query complexities that scale polynomially with both the dimension d and the error tolerance ε(e.g., \tilde\mathcalO(d/ε)). This mirrors the limitations of unadjusted Langevin algorithm, where standard first-order score solvers lack access to zeroth-order density information, precluding natural error-correction mechanisms and thus preventing the fast \ln(1/ε) convergence attainable by Metropolis-adjusted methods. In this paper, we develop an improved generative modeling method by introducing \ourmethod (Quantized Transition Diffusion), a framework that reformulates continuous diffusion into a discrete generation problem through spatial quantization and the parameterization of zeroth-order information (e.g., density ratios). To sample from this discrete target, we propose a truncated uniformization algorithm that simulates the underlying continuous-time Markov chain of the discrete diffusion process without discretization error, while eliminating the restrictive bounded-score assumption required by prior uniformization-based approaches. Consequently, \ourmethod attains ε-accuracy in total variation distance with a query complexity of \mathcalO(d \ln^2(d/ε)), yielding a notable improvement in ε-dependence compared to existing continuous diffusion samplers. Crucially, our analysis capitalizes on a novel proof technique based on the infinitesimal chain rule of KL divergence, providing a fresh perspective on unifying continuous and discrete diffusion paradigms.
@article{huang2026almost,abbr={COLT},bibtex_show={true},title={Almost linear convergence under minimal score assumptions: Quantized transition diffusion},author={Huang, Xunpeng and Lin, Yingyu and Kuang, Nikki Lijing and Dong, Hanze and Zou, Difan and Ma, Yian and Zhang, Tong},journal={Annual Conference on Learning Theory (COLT)},year={2026},html={https://arxiv.org/pdf/2505.21892},selected={true}}
NeurIPS
Reverse Transition Kernel: A Flexible Framework to Accelerate Diffusion Inference
Xunpeng Huang,
Difan Zou,
Hanze Dong,
Yi Zhang,
Yian Ma,
and Tong Zhang
In Annual Conference on Neural Information Processing Systems (NeurIPS) 2024, Best paper of ICML Workshop on Structured Probabilistic Inference & Generative Modeling
2024
To generate data from trained diffusion models, most inference algorithms, such as DDPM, DDIM, and other variants, rely on discretizing the reverse SDEs or their equivalent ODEs. In this paper, we view such approaches as decomposing the entire denoising diffusion process into several segments, each corresponding to a reverse transition kernel (RTK) sampling subproblem. Specifically, DDPM uses a Gaussian approximation for the RTK, resulting in low per-subproblem complexity but requiring a large number of segments (i.e., subproblems), which is conjectured to be inefficient. To address this, we develop a general RTK framework that enables a more balanced subproblem decomposition, resulting in \tildeO(1) subproblems, each with strongly log-concave targets. We then propose leveraging two fast sampling algorithms, the Metropolis-Adjusted Langevin Algorithm (MALA) and Underdamped Langevin Dynamics (ULD), for solving these strongly log-concave subproblems. This gives rise to the RTK-MALA and RTK-ULD algorithms for diffusion inference. In theory, we further develop the convergence guarantees for RTK-MALA and RTK-ULD in total variation (TV) distance: RTK-ULD can achieve εtarget error within \tildeO(d^1/2ε^-1) under mild conditions, and RTK-MALA enjoys a O(d^2\log(d/ε)) convergence rate under slightly stricter conditions. These theoretical results surpass the state-of-the-art convergence rates for diffusion inference and are well supported by numerical experiments.
@inproceedings{huang2024reverse,abbr={NeurIPS},bibtex_show={true},title={Reverse Transition Kernel: A Flexible Framework to Accelerate Diffusion Inference},author={Huang, Xunpeng and Zou, Difan and Dong, Hanze and Zhang, Yi and Ma, Yian and Zhang, Tong},booktitle={Annual Conference on Neural Information Processing Systems (NeurIPS) 2024, Best paper of ICML Workshop on Structured Probabilistic Inference & Generative Modeling},year={2024},html={https://arxiv.org/abs/2401.06325},selected={true}}
COLT
Faster Sampling without Isoperimetry via Diffusion-based Monte Carlo
To sample from a general target distribution p∗∝e−f∗ beyond the isoperimetric condition, Huang et al. (2023) proposed to perform sampling through reverse diffusion, giving rise to Diffusion-based Monte Carlo (DMC). Specifically, DMC follows the reverse SDE of a diffusion process that transforms the target distribution to the standard Gaussian, utilizing a non-parametric score estimation. However, the original DMC algorithm encountered high gradient complexity, resulting in an exponential dependency on the error tolerance ϵ of the obtained samples. In this paper, we demonstrate that the high complexity of DMC originates from its redundant design of score estimation, and proposed a more efficient algorithm, called RS-DMC, based on a novel recursive score estimation method. In particular, we first divide the entire diffusion process into multiple segments and then formulate the score estimation step (at any time step) as a series of interconnected mean estimation and sampling subproblems accordingly, which are correlated in a recursive manner. Importantly, we show that with a proper design of the segment decomposition, all sampling subproblems will only need to tackle a strongly log-concave distribution, which can be very efficient to solve using the Langevin-based samplers with a provably rapid convergence rate. As a result, we prove that the gradient complexity of RS-DMC only has a quasi-polynomial dependency on ϵ, which significantly improves exponential gradient complexity in Huang et al. (2023). Furthermore, under commonly used dissipative conditions, our algorithm is provably much faster than the popular Langevin-based algorithms. Our algorithm design and theoretical framework illuminate a novel direction for addressing sampling problems, which could be of broader applicability in the community.
@inproceedings{huang2024faster,abbr={COLT},bibtex_show={true},title={Faster Sampling without Isoperimetry via Diffusion-based Monte Carlo},author={Huang, Xunpeng and Zou, Difan and Dong, Hanze and Ma, Yian and Zhang, Tong},booktitle={Annual Conference on Learning Theory (COLT)},year={2024},html={https://arxiv.org/abs/2401.06325},selected={true}}
ICML
Faster Sampling via Stochastic Gradient Proximal Sampler
Stochastic gradients have been widely integrated into Langevin-based methods to improve their scalability and efficiency in solving large-scale sampling problems. However, the proximal sampler, which exhibits much faster convergence than Langevin-based algorithms in the deterministic setting \citelee2021structured, has yet to be explored in its stochastic variants. In this paper, we study the Stochastic Proximal Samplers (SPS) for sampling from non-log-concave distributions. We first establish a general framework for implementing stochastic proximal samplers and establish the convergence theory accordingly. We show that the convergence to the target distribution can be guaranteed as long as the second moment of the algorithm trajectory is bounded and restricted Gaussian oracles can be well approximated. We then provide two implementable variants based on Stochastic gradient Langevin dynamics (SGLD) and Metropolis-adjusted Langevin algorithm (MALA), giving rise to SPS-SGLD and SPS-MALA. We further show that SPS-SGLD and SPS-MALA can achieve ε-sampling error in total variation (TV) distance within \tilde\mathcalO(dε^-2) and \tilde\mathcalO(d^1/2ε^-2) gradient complexities, which outperform the best-known result by at least an \tilde\mathcalO(d^1/3) factor. This enhancement in performance is corroborated by our empirical studies on synthetic data with various dimensions, demonstrating the efficiency of our proposed algorithm.
@inproceedings{huang2024sps,abbr={ICML},bibtex_show={true},title={Faster Sampling via Stochastic Gradient Proximal Sampler},author={Huang, Xunpeng and Zou, Difan and Ma, Yian and Dong, Hanze and Zhang, Tong},booktitle={International Conference on Machine Learning (ICML)},year={2024},html={https://proceedings.mlr.press/v235/huang24aj.html},selected={true}}
The efficacy of modern generative models is commonly contingent upon the precision of score estimation along the diffusion path, with a focus on diffusion models and their ability to generate high-quality data samples. This study delves into the application of reverse diffusion to Monte Carlo sampling. It is shown that score estimation can be transformed into a mean estimation problem via the decomposition of the transition kernel. By estimating the mean of the posterior distribution, we derive a novel Monte Carlo sampling algorithm from the reverse diffusion process, which is distinct from traditional Markov Chain Monte Carlo (MCMC) methods. We calculate the error requirements and sample size for the posterior distribution, and use the result to derive an algorithm that can approximate the target distribution to any desired accuracy. Additionally, by estimating the log-Sobolev constant of the posterior distribution, we show under suitable conditions the problem of sampling from the posterior can be easier than direct sampling from the target distribution using traditional MCMC techniques. For Gaussian mixture models, we demonstrate that the new algorithm achieves significant improvement over the traditional Langevin-style MCMC sampling methods both theoretically and practically. Our algorithm offers a new perspective and solution beyond classical MCMC algorithms for challenging complex distributions.
@inproceedings{huang2024reversf,abbr={ICLR},bibtex_show={true},title={Reverse Diffusion Monte Carlo},author={Huang, Xunpeng and Dong, Hanze and HAO, Yifan and Ma, Yian and Zhang, Tong},booktitle={International Conference on Learning Representations (ICLR)},year={2024},html={https://openreview.net/forum?id=kIPEyMSdFV},selected={true}}
