Weighted Adjoint Matching: Finetuning Flow Matching Models with Time-Dependent Control Cost
stat.ML, cs.LG
Submitted: 2026-10-14
Updated: 2026-10-14
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Certifying Some Distributional Robustness with Principled Adversarial Training
- Avoiding mode collapse in diffusion models fine-tuned with reinforcement learning
- Understanding Reinforcement Learning-Based Fine-Tuning of Diffusion Models: A Tutorial and Review
- Fine-Tuning of Continuous-Time Diffusion Models as Entropy-Regularized Control
- Diffusion Models Generate Images Like Painters: an Analytical Theory of Outline First, Details Later
- Improved techniques for fine-tuning flow models via adjoint matching: a deterministic control pipeline
- Differentiating through Stochastic Differential Equations: A Primer
- The tractability landscape of diffusion alignment: regularization, rewards, and computational primitives
Related papers
- Behavior of prediction performance metrics with rare events
- Optimal Estimation of Generic Dynamics by Path-Dependent Neural Jump ODEs
- A Posterior-Dynamics Framework for Imaging Inverse Problems with Pretrained Diffusion Priors
- Scalable Posterior Uncertainty for Flexible Density-Based Clustering
- A Transport Geometry of Benign Overfitting
- Efficient quadratic entropy with distance sketches