mlsanityData to 14 Oct 2026

Learning ω-Powered Conditional Distributions with Importance-Weighted Flow Matching

arXiv 2610.16419 ↗First posted 12 Oct 2026cs.LG cs.CVcited by 0

On the map: Diffusion models → Flow matching, time series and science also fits: Diffusion and flow policies for robotics

Abstract (arXiv)

Classifier-Free Guidance (CFG) with guidance scale γ is commonly motivated by the ω-powered conditional distribution obtained by setting ω=γ, but CFG does not, in general, sample from this distribution exactly. We study whether the powered target can instead be learned explicitly. We introduce Importance-Weighted Conditional Flow Matching (IW-CFM), which reweights conditional flow-matching examples using density ratios obtained from frozen conditional and unconditional flow models. With exact normalized weights, the resulting objective has the same gradient as flow matching on the powered target; in practice, we use self-normalized estimation and a regularized empirical approximation for high-dimensional data. On synthetic distributions, IW-CFM recovers analytical powered targets for which numerically matched CFG, γ=ω, exhibits a visible mismatch. On ImageNet, an amortized lightweight corrector achieves better alignment than fixed- and matched-guidance CFG with the corresponding weighted and filtered reference distributions. We additionally present an exploratory CLIP analysis of powered-density rankings on real ImageNet data. Aggregate shifts across selected prompt-specific contrasts coexist with heterogeneous classwise responses, including nonmonotonic trajectories. These results provide a practical route for studying and approximating model-defined powered conditional distributions.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Flow matching, time series and science, one of 14 directions in Diffusion models. Builds on 34 of its 42 references that are in the map. First posted on arXiv 12 Oct 2026.

Builds on

34 of 42 references.

Closest ancestors · 4

Same direction, newest first

  1. 2023Guided Flows for Generative Modeling and Decision Making22 Nov
  2. 2023Flow Matching in Latent Space17 Jul
  3. 2022Flow Matching for Generative Modeling6 Oct
  4. 2022Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow7 Sep

Other references · 30

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
  2. 2020Denoising Diffusion Probabilistic Modelscited by 21,076
  3. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,863
  4. 2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 9,122
  5. 2021Diffusion Models Beat GANs on Image Synthesiscited by 8,255
  6. 2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,058
  7. 2022Classifier-Free Diffusion Guidancecited by 5,847
  8. 2017On Calibration of Modern Neural Networkscited by 4,960
  9. 2018Neural Ordinary Differential Equationscited by 3,779
  10. 2022Exploring CLIP for Assessing the Look and Feel of Imagescited by 991
  11. 2019Improved Precision and Recall Metric for Assessing Generative Modelscited by 944
  12. 2023Stochastic Interpolants: A Unifying Framework for Flows and Diffusionscited by 939
18 more
  1. 2024SiT: Exploring Flow and Diffusion-based Generative Models with Scalable Interpolant Transformerscited by 737
  2. 2022On Distillation of Guided Diffusion Modelscited by 725
  3. 2020Reliable Fidelity and Diversity Metrics for Generative Modelscited by 434
  4. 2024Guiding a Diffusion Model with a Bad Version of Itselfcited by 299
  5. 2023Blind Image Quality Assessment via Vision-Language Correspondence: A Multitask Learning Perspectivecited by 280
  6. 2023Exposing flaws of generative model evaluation metrics and their unfair treatment of diffusion modelscited by 205
  7. 2024CFG++: Manifold-constrained Classifier Free Guidance for Diffusion Modelscited by 149
  8. 2023On Memorization in Diffusion Modelscited by 125
  9. 2025Feynman-Kac Correctors in Diffusion: Annealing, Guidance, and Product of Expertscited by 96
  10. 2024Classifier-Free Guidance is a Predictor-Correctorcited by 84
  11. 2017On Nesting Monte Carlo Estimatorscited by 83
  12. 2025Energy-Weighted Flow Matching for Offline Reinforcement Learningcited by 73
  13. 2024What does guidance do? A fine-grained analysis in a simple settingcited by 48
  14. 2025Debiasing Guidance for Discrete Diffusion with Sequential Monte Carlocited by 31
  15. 2024Diffusion Models as Cartoonists: The Curious Case of High Density Regionscited by 21
  16. 2025Learn to Guide Your Diffusion Modelcited by 14
  17. 2025REG: Rectified Gradient Guidance for Conditional Diffusion Modelscited by 8
  18. 2025Efficient Distillation of Classifier-Free Guidance using Adapterscited by 6

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Classifier-Free Guidance in Flow Matching: Non-Autonomous Potentials, Overshoot, and Posterior-Mean Control
  2. 2026P-Guide: Parameter-Efficient Prior Steering for Single-Pass CFG Inference
  3. 2025Weighted Conditional Flow Matching
  4. 2025On the Guidance of Flow Matching
  5. 2026Improving Classifier-Free Guidance of Flow Matching via Manifold Projection
  6. 2025Classifier-Free Guidance: From High-Dimensional Analysis to Generalized Guidance Forms
  7. 2026Information-Theoretic Classifier-Free Guidance with Adaptive Schedule Optimization
  8. 2026TFTF: Training-Free Targeted Flow for Conditional Sampling