Inference-Time Alignment for Diffusion Models via Variationally Stable Doob's Matching
stat.ML, cs.LG, cs.NA, math.NA, math.OC, math.ST, stat.TH
Submitted: 2026-01-10
Updated: 2026-11-02
Project page: https://chewisinho.github.io/main.pdf
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- DogFit: Domain-guided Fine-tuning for Efficient Transfer Learning of Diffusion Models
- Convergence of Deterministic and Stochastic Diffusion-Model Samplers: A Simple Analysis in Wasserstein Distance
- Deep conditional distribution learning via conditional F"ollmer flow
- Provable Diffusion Posterior Sampling for Bayesian Inversion
- Solving Inverse Problems via Diffusion-Based Priors: An Approximation-Free Ensemble Sampling Approach
- Iterative Importance Fine-tuning of Diffusion Models
- Nonlinear Assimilation via Score-based Sequential Langevin Sampling
- Characteristic Learning for Provable One Step Generation
- Unveil Conditional Diffusion Models with Classifier-free Guidance: A Sharp Statistical Theory
- Stochastic Control for Fine-tuning Diffusion Models: Optimality, Regularity, and Convergence
- Denoising diffusion probabilistic models are optimally adaptive to unknown low dimensionality
- Towards a unified framework for guided diffusion models
- Non-asymptotic error bounds for probability flow ODEs under weak log-concavity
- Aligning Text-to-Image Models using Human Feedback
- Is RL fine-tuning harder than regression? A PDE learning approach for diffusion models
- Iterative Tilting for Diffusion Fine-Tuning
- Linear Convergence of Diffusion Models Under the Manifold Hypothesis
- DriftLite: Lightweight Drift Control for Inference-Time Scaling of Diffusion Models
- Test-time scaling of diffusions with flow maps
- Differentiable Neural Networks with RePU Activation: with Applications to Score Estimation and Isotonic Regression
Related papers
- Uniform-in-Time Weak Propagation-of-Chaos in Shallow Neural Networks
- Diffusion Models Observe Only Gradients: A Geometric Perspective on Score Matching Errors
- Quantifying Error Propagation and Model Collapse in Diffusion Models
- Behavior of prediction performance metrics with rare events
- Statistical Inference and Learning for Shapley Additive Explanations (SHAP)
- Control Variate Design for Score-Function Gradient Estimators in Importance-Weighted Variational Inference