Flow-Based Generative Modeling

Overview

Flow-based generative modeling represents generation as a transformation of probability mass, often by integrating a continuous-time velocity field from noise to data or composing invertible normalizing-flow layers. Flow matching and rectified flow train velocity fields by regression along selected paths; optimal-transport couplings, stochastic interpolants, discrete Markov flows, and flow maps extend the toolkit to different data types and sampling budgets. Researchers apply these methods to images, language, molecules, physical fields, and robot actions, assessing fidelity and diversity alongside likelihood, constraint satisfaction, calibration, and runtime. A central challenge is achieving one- or few-step generation without sacrificing quality or mode coverage, while improving path design, robustness, geometric and physical consistency, and theoretical guarantees.

Latest papers

CardsList

619 papers57 papers in the last four weeks, down 57% on the four weeks before. 0.4% of all new papers.

  1. Flow Matching from Non-Gaussian Sources: A Theory of Efficient Sampling

    Nov 19, 2026Yuchen Zhou, Lili Zheng, Yuxin Chen +1Flow MatchingFlow-Based Generative Modeling

  2. SpecFlow: Flow Matching in the Acceleration Spectrum for Multi-Agent Trajectory Forecasting

    Nov 19, 2026Ziyang Xu, Peng Han, Hai ZhouMulti-Agent Trajectory PredictionMultimodal Trajectory Prediction

  3. One-to-Many: Diverse and Bidirectional Image-to-Image Translation via Rectified Flow

    Nov 19, 2026Sho Shimotsumagari, Seiichi Uchida, Shumpei TakezakiImage-to-Image TranslationFlow-Based Generative Modeling

  4. Why Flat Drafts Stay Flat: Region-Mean Latents Govern Texture in Flow-Based Image Generation

    Nov 19, 2026Zihan Gu, Jiangcan Jia, Hua Zhang +1Flow-Based Generative ModelingControllable Image Generation

  5. LIFT: Latent Flow Matching for Probabilistic Cell Fate Dynamics with Single Cell Snapshots

    Nov 19, 2026Bowen Yang, Junda Ying, Yuxuan Wang +2Latent Dynamics ModelingConditional Generative Modeling

  6. TrajCal: Trajectory-Calibrated Time Schedules for Few-Step Flow Sampling

    Nov 19, 2026Junchi Zhang, Zihan Gu, Jiangcan Jia +3Flow-Based Generative ModelingFlow Matching

  7. FamilySpace: From Emergent Chemical Semantics to Controllable Crystal Generation

    Nov 19, 2026Haowei Hua, Irwin King, Wanyu LinCrystal Structure GenerationConstrained Generative Modeling

  8. Gaspard: Expressive unified speech and gesture synthesis with continuous autoregressive flow-matching and decoupled classifier-free guidance

    Nov 18, 2026Téo Guichoux, Théodor Lemerle, Catherine Pelachaud +1Multimodal GenerationUnified Multimodal Models

  9. Optimal Control of RKHS Diffeomorphic Flows for Classification and Generative Modeling

    Nov 17, 2026Christopher Chukwuemeka, Andreas MangFlow-Based Generative ModelingDensity Estimation

  10. Correcting Flows with Marginal Matching

    Nov 17, 2026Meital Bojan, Nadav Bojan Sellam, Ev Zisselman +1Flow MatchingFlow-Based Generative Modeling

  11. The Limits of Closed-Form Couplings for Few-Step Generation

    Nov 16, 2026Tobias Höppe, Stefan Bauer, Qiang Liu +2Flow-Based Generative ModelingFlow Matching

  12. Multi-Rate Flow Matching for Point Cloud Upsampling

    Nov 16, 2026Jingkai Shang, Edgar Torres, Duy NguyenPoint Cloud LearningFlow Matching

  13. Explicit Regularization for Flow Matching

    Nov 14, 2026Lars Ruthotto, Eran TreisterFlow MatchingFlow-Based Generative Modeling

  14. Vessel2Brain: Edge-Aware Multi-Scale Flow Matching for 3D Brain Image Generation from Vessel Masks

    Nov 13, 2026Zhi Chen, Tianyang Zhang, Theodoros N. Arvanitis +1Medical Image SynthesisBrain MRI Synthesis

  15. Learning Anisotropic Prediction Targets for Flow Models

    Nov 13, 2026Rayhan Zirvi, Hansheng Chen, Gordon Wetzstein +1Flow-Based Generative ModelingFlow Matching

  16. Finite speed Ornstein-Uhlenbeck-Kac processes and their applications in generative modeling

    Nov 13, 2026Richard Duong, Janis Auffenberg, Gabriele SteidlMarkov ModelsFlow-Based Generative Modeling

  17. FréchetFlow: Flow Matching for Non-Euclidean Data

    Nov 11, 2026Yidong Zhou, Muqing Cui, Hans-Georg MüllerGenerative ModelingFlow-Based Generative Modeling

  18. AdjointFlow: Accelerating Scientific Inverse Problems with Physics-Based Transport

    Nov 10, 2026Ethan Hanold, Nicholas M. Boffi, Qin Li +1Inverse ProblemsPhysics-Informed Generative Modeling

  19. The balance between feature learning and collapse in dynamical systems-based generative models

    Nov 9, 2026Julian Brandon, Bruno Loureiro, N Alex Cayco-Gajic +1Generative ModelingMode Collapse

  20. GenRM-Flow: Generators are Process-aware Reward Models in Flow Matching

    Nov 8, 2026Siming Fu, Ruizhe He, Zheming Fu +7Reward Modeling for Video GenerationVideo Reward Modeling

  21. EbbFlow: PINN-Inspired Progressive Curriculum for MeanFlow Training

    Nov 8, 2026Mengfan Yuan, Xunheng Kuang, Jianbin Chen +1Flow-Based Generative ModelingFlow Matching

  22. Constrained Look-ahead Guidance for Interference-Aware Flow Editing

    Nov 7, 2026Doudou Zhang, Qi ChenFlow-Based Image EditingImage Editing

  23. CoMoFlow: Controllable Protein Conformational Ensemble Generation via Explicit Collective-Mode Modeling

    Nov 6, 2026Yusong Wang, Zheyuan Gu, Ailiang Lin +5Protein Conformational Ensemble GenerationProtein Structure Generation

  24. Continued Training of Continuous Language Flows with Token Identity Squared Supervision

    Nov 5, 2026Jiawei Jiang, Zhaoxiang Zang, Zhengtao YuDiffusion Language Model TrainingContinuous Diffusion Language Models

  25. Generator-Aligned Reward Modeling with Target-Aware Bounded Optimization for Visual Generation

    Nov 5, 2026Ziqiao Weng, Yuanzhi Liang, Haibin Huang +2Reward ModelingControllable Video Generation

  26. Energy-Guided Transport for Projection-Free Physics-Informed Flow Matching

    Nov 4, 2026Giuliano Galadini, Aurelio Uncini, Danilo ComminielloPhysics-Informed Generative ModelingConstrained Generative Modeling