Latest Flow Matching Research Papers
The newest Flow Matching papers from across the field — arXiv, NeurIPS, CVPR, Nature, and more — refreshed daily and ranked by relevance. Distill AI tracks Flow Matching so you don’t have to: get the standout work delivered to your inbox every morning, with 2-sentence summaries and the option to chat with any paper.
Get the latest Flow Matching papers in your inbox — free →Recent papers
- Model-Aware Schedules Improve Generation via Fiberwise Optimal TransportLuyi Jia, Boyan Zhang, Yilun Liu, Steffen Rulands · arXiv · Sep 10, 2026
Diffusion and flow-matching schedules control the signal and noise coefficients that mix data and noise along affine probability paths. Minimizing a kinetic action defined on coefficient paths, motivated by optimal transport, helps explain …
- LoRA-TSD: Tangent-Space Spectral Descent for LoRA via Muon-Style UpdatesDmitrii Andriianov, Andrey Veprikov, Aleksandr Beznosikov · arXiv · Sep 2, 2026
Low-rank adaptation (LoRA) is the standard way to fine-tune large models, yet when its two factors are trained independently, the update ignores the geometry of the low-rank weight change it induces. We introduce LoRA-TSD, an optimizer that…
- Language-Informed Flow Matching for Trend-Guided Structure-Based 3D Molecular GenerationTianyu Gao, Zhikai Su, Jiashu Li, Wenjun Gao et al. · arXiv · Aug 31, 2026
Structure-based drug design (SBDD) requires ligands that satisfy both 3D target affinity and 1D chemical validity. Existing controllable generation methods often rely on task-specific fine-tuning or externally imposed sampling-time guidance…
- How do flow matching models memorize and generalize in sample data subspacesWeiguo Gao, Ming Li · npj Artificial Intelligence · Aug 25, 2026
Abstract Real-world data often lie near a low-dimensional structure embedded in a high-dimensional space. In practice, we observe only finitely many samples, which define a sample data subspace that supports tasks such as dimensionality red…
- On the Fragility of Self-Improving Agents: Variance, Task Order, and UnderspecificationQinyuan Ye, Yu Li, Yada Pruksachatkun, Jiaxin Zhang et al. · arXiv · Aug 18, 2026
Memory-based self-improving agents--those that learn from an online stream of tasks and improve over time by maintaining a textual memory bank--have shown great promise in recent literature. However, the reliability aspects of these methods…
- Composing Flow-Matching Energies with Known Physics: Generation, OOD Detection, and Inversion on PDE FieldsYixuan Sun, Anirban Samaddar, Sandeep Madireddy · arXiv · Aug 18, 2026
Probabilistic modeling of physical fields benefits from both a data-driven prior and known physical structure such as the governing equations. Energy-based models (EBMs) are a natural fit since energies compose additively, which enables aug…
- Collaborative Probabilistic Wind Power Forecasting Using Adaptive Flow Matching Enhanced Neural ProcessJiabei Liu, Zheren Zhu, Le Yao, Jiusun Zeng · International Journal of Ad... · Aug 13, 2026
ABSTRACT With the large‐scale integration of wind power, its inherent intermittency and uncertainty pose significant challenges to power system operation. Existing probabilistic forecasting methods often struggle with capturing spatiotempor…
- Efficient identification of critical regions via Flow Matching–based Monte Carlo initializationQian-Rui Lee, Daw-Wei Wang · Machine Learning Science an... · Aug 6, 2026
Abstract Markov chain Monte Carlo (MCMC) is a standard tool for studying many-body systems, but its practical cost can become substantial, especially when simulations must be repeated across temperatures and lattice sizes or near transition…
- Multimodal Spatiotemporal Atmospheric Data Assimilation with Latent Flow-matchingDibyajyoti Chakraborty, Romit Maulik · arXiv · Aug 5, 2026
Data assimilation (DA) uses Bayesian inference to update the state of a numerical forecast model with observed data. In this study, we propose a fundamentally different, unified approach to atmospheric data assimilation. We use latent video…
- Yield-aware generative inverse design of anti-reflection coatings via optimal-transport flow matching with a conditional-value-at-risk objectiveErfan Ghapanvari · Scientific Reports · Aug 5, 2026
Abstract Robustness to deposition error is conventionally treated as a post-hoc Monte-Carlo check in multilayer anti-reflection (AR) coating design: a nominally optimal stack is found first, and its manufacturing tolerance is assessed after…
- Computational and Statistical Guarantees of the \textit{c}-Rectified flowLeda Wang, Zhehao Xu, Qiang Liu, Harrison H. Zhou · arXiv · Aug 3, 2026
Recently, rectified flow has emerged as a fundamental framework for large-scale image generation, powering state-of-the-art systems such as FLUX.1 and Stable Diffusion 3. Despite its remarkable empirical success, the computational and stati…
- Certified Parallel-in-Time Sinkhorn for Dynamic Entropic Optimal TransportXinyang Wen · arXiv · Jul 27, 2026
Dynamic applications, including optimal-transport Flow Matching, repeatedly solve related entropic optimal transport problems, yet conventional distributed Sinkhorn processes frames sequentially and synchronizes after every iteration. We pr…
- When Can You Correct Distribution Drift in Temporal Graph Generation? A Sharpening--Drift Tension and an Impossibility for Observation-Based CorrectionTianpeng Li, Xuan Guo, Wenjun Wang, Wang Zhang et al. · arXiv · Jul 27, 2026
Generative models of temporal graphs are trained on one stretch of an evolving network and deployed on the next, and they degrade badly in the gap. We show this degradation is derivable, general, and not fixable from observations. The maske…
- ROMS-IMLE: A Minimalist Approach to Competitive Single-Step Generative ModellingChirag Vashist, Ke Li · arXiv · Jul 21, 2026
Generative models have undergone many generations of evolution, from VAEs/GANs to diffusion/flow matching. Along the way, the underlying techniques have become more complicated and various beliefs about what drives strong empirical performa…
- A Shortcut to Statistically Steady-State Turbulence with Flow MatchingGianluca Galletti, Gerald Gutenbrunner, William Hornsby, Lorenzo Zanisi et al. · arXiv · Jul 14, 2026
Many nonlinear physical systems exhibit an initial transient phase in which perturbations grow before nonlinear interactions lead to a statistically steady state. While this saturated regime is of primary interest, direct numerical simulati…
- Learning Manifold Data with Flow MatchingSophia Pi, Mingcheng Lu, Jerry Yao-Chieh Hu, Maojiang Su et al. · SPIGM @ ICML Poster · May 30, 2026
We study flow-matching transformers when data lie on a low-dimensional manifold. Our key insight is a flow decomposition that splits motion along the manifold from motion off the manifold. The scheme works for first- and higher-order flow m…
- Learning Manifold Data with Flow MatchingSophia Pi, Mingcheng Lu, Jerry Yao-Chieh Hu, Maojiang Su et al. · ICML 2026 FoGen Workshop Poster · May 26, 2026
We study flow-matching transformers when data lie on a low-dimensional manifold. Our key insight is a flow decomposition that splits motion along the manifold from motion off the manifold. The scheme works for first- and higher-order flow m…
- Learning Manifold Data with Flow MatchingSophia Pi, Mingcheng Lu, Maojiang Su, Weimin Wu et al. · ICML 2026 regular · Apr 30, 2026
We study flow-matching transformers when data lie on a low-dimensional manifold. Our key insight is a flow decomposition that splits motion along the manifold from motion off the manifold. The scheme works for first and higher-order flow …
- Low-Pass Flow MatchingFrancesco M. Ruscio, T. Konstantin Rusch · ICLR 2026 DeLTa Workshop Poster · Mar 3, 2026
Flow Matching typically relies on white noise sources, a choice often misaligned with the power spectra of natural data, which tend to decay with frequency. To address this, we introduce $\textbf{Low-Pass Flow Matching}$, a variant of Flow …
- Flow Matching Policy GradientsDavid McAllister, Songwei Ge, Brent Yi, Chung Min Kim et al. · ICLR 2026 Poster · Jan 26, 2026
Flow-based generative models, including diffusion models, excel at modeling continuous distributions in high-dimensional spaces. In this work, we introduce Flow Policy Optimization (FPO), a simple on-policy reinforcement learning algorithm …
- Topological Flow MatchingKacper Wyrwal, Ismail Ilkan Ceylan, Alexander Tong · ICLR 2026 Poster · Jan 26, 2026
Flow matching is a powerful generative modeling framework, valued for its simplicity and strong empirical performance. However, its standard formulation treats signals on structured spaces---such as fMRI data on brain graphs---as points in …
- Active Flow MatchingYashvir Singh Grewal, Edwin V. Bonilla, Thang D Bui · AIML-CEB 2025 Oral · Nov 12, 2025
Discrete diffusion and flow matching excel at capturing epistatic structure in protein fitness landscapes through parallel, iterative refinement. However, their implicit nature—sampling via learned dynamics without tractable densities—preve…
- Federated Flow MatchingZifan Wang, Anqi Dong, Mahmoud Selim, Michael M. Zavlanos et al. · Submitted to ICLR 2026 · Sep 19, 2025
Data today is decentralized, generated and stored across devices and institutions where privacy, ownership, and regulation prevent centralization. This motivates the need to train generative models directly from distributed data locally wit…
- Coefficients-Preserving Sampling for Reinforcement Learning with Flow MatchingFeng Wang, Zihao Yu · arXiv.org · Sep 7, 2025
Reinforcement Learning (RL) has recently emerged as a powerful technique for improving image and video generation in Diffusion and Flow Matching models, specifically for enhancing output quality and alignment with prompts. A critical step f…
- Flow Matching Policy GradientsDavid McAllister, Songwei Ge, Brent Yi, C. Kim et al. · arXiv.org · Jul 28, 2025
Flow-based generative models, including diffusion models, excel at modeling continuous distributions in high-dimensional spaces. In this work, we introduce Flow Policy Optimization (FPO), a simple on-policy reinforcement learning algorithm …
- La-Proteina: Atomistic Protein Generation via Partially Latent Flow MatchingTomas Geffner, Kieran Didi, Zhonglin Cao, Danny Reidenbach et al. · arXiv.org · Jul 13, 2025
Recently, many generative models for de novo protein structure design have emerged. Yet, only few tackle the difficult task of directly generating fully atomistic structures jointly with the underlying amino acid sequence. This is challengi…
- FLUX.1 Kontext: Flow Matching for In-Context Image Generation and Editing in Latent SpaceBlack Forest Labs, Stephen Batifol, A. Blattmann, Frederic Boesel et al. · arXiv.org · Jun 17, 2025
We present evaluation results for FLUX.1 Kontext, a generative flow matching model that unifies image generation and editing. The model generates novel output views by incorporating semantic context from text and image inputs. Using a simpl…
- Physics-Constrained Flow Matching: Sampling Generative Models with Hard ConstraintsUtkarsh Utkarsh, Pengfei Cai, Alan Edelman, Rafael Gómez-Bombarelli et al. · Neural Information Processing Systems · Jun 4, 2025
Deep generative models have recently been applied to physical systems governed by partial differential equations (PDEs), offering scalable simulation and uncertainty-aware inference. However, enforcing physical constraints, such as conserva…
- On the Closed-Form of Flow Matching: Generalization Does Not Arise from Target StochasticityQuentin Bertrand, Anne Gagneux, Mathurin Massias, R'emi Emonet · Advances in Neural Information Processing Systems 38 · Jun 4, 2025
Modern deep generative models can now produce high-quality synthetic samples that are often indistinguishable from real training data. A growing body of research aims to understand why recent methods, such as diffusion and flow matching tec…
- ReinFlow: Fine-tuning Flow Matching Policy with Online Reinforcement LearningTonghe Zhang, Chao Yu, Sichang Su, Yu Wang · Neural Information Processing Systems · May 28, 2025
We propose ReinFlow, a simple yet effective online reinforcement learning (RL) framework that fine-tunes a family of flow matching policies for continuous robotic control. Derived from rigorous RL theory, ReinFlow injects learnable noise in…