Score-Based Generative Modeling

Overview

Score-based generative modeling learns the gradient of a data distribution’s log density, often at multiple noise levels using denoising score matching. Generation then follows learned scores through reverse-time stochastic differential equations, probability-flow ordinary differential equations, or Langevin correctors; extensions handle manifolds, discrete data, and scientific inverse problems. Methods are evaluated on image-generation benchmarks and tasks such as MRI reconstruction, simulation-based inference, and data assimilation, where fidelity, calibration, and sampling cost all matter. A central challenge is connecting score-estimation accuracy to reliable samples: low-noise errors, discretization, initialization, and data geometry can all affect results, while stronger smoothing may reduce memorization but alter fidelity and diversity.

Latest papers

122 papers10 papers in the last four weeks, down 41% on the four weeks before. 0.1% of all new papers.

Oct 2, 2024

Robustness and Structure Preservation in Flow-Based Generative Models via Wasserstein Path-Space Divergences

We introduce a novel Wasserstein-1 (W1W_1) path-space divergence for stochastic and deterministic dynamics and establish a Wasserstein Uncertainty Propagation (WUP) theorem that bounds the W1W_1 distance between terminal distributions by the proposed divergence, equivalently characterized by a weighted L2L^2 discrepancy between the underlying drifts and the W1W_1 distance between their initial measures. A key ingredient is a probabilistic framework combining adjoint Feynman-Kac representations with synchronous coupling (and reflection coupling on bounded domains), yielding Wasserstein stability estimates beyond existing PDE- and Girsanov-based approaches. The framework accommodates time-varying and possibly degenerate diffusion coefficients, empirical and singular measures, and remains valid in the deterministic limit of flow matching. Unlike KL-based uncertainty quantification bounds, it does not require absolute continuity of path measures and therefore remains well-defined in singular settings. As consequences of the WUP theorem, we derive W1W_1 robustness and generalization bounds for score-based generative models and flow matching at both population and finite-sample levels. We further specialize the framework to group-symmetric targets, providing the first error analysis of equivariant flow-based models and the first quantitative comparison between data augmentation and equivariant inductive bias. Our analysis identifies a symmetry-aware Wasserstein path-space divergence that quantifies the model-form error induced by non-equivariant parametrizations. We prove that this error cannot be removed by additional data or training and vanishes only under equivariant architectures, establishing a precise theoretical advantage of equivariant inductive bias over data augmentation. Numerical experiments on group-symmetric Gaussian mixtures corroborate the theory.
May 22, 2024

Adaptive Momentum Benefits Score-based Generative Models

Score-based Generative Models (SGMs) have achieved remarkable success in generation tasks by establishing a diffusion process that gradually perturbs real data into Gaussian noise and then learning a reverse process to reconstruct the data from noise. However, conventional denoising methods, such as Langevin dynamics and numerical stochastic differential equation solvers, rely on randomness but suffer from slow generation speeds due to the high number of score function evaluations. To address this, we propose a plug-and-play Markov adaptive momentum sampling method, inspired by the strong connection between the SGMs' sampling process and stochastic gradient descent optimization. Our approach accelerates the transformation process without introducing additional hyperparameters. Theoretically, we establish a rigorous foundation by analyzing the relationship between the distribution of samples generated by our method and the original data distribution. This provides a novel perspective on the analysis of sampling distributions. Empirically, our method demonstrates superior efficiency, generating high-quality images/graphs in significantly fewer sampling steps, achieving a speedup of 2x to 5x while maintaining or surpassing baseline performance on various image and graph generation benchmarks. These results underscore the effectiveness of adaptive momentum in enhancing the efficiency of SGMs.