CS Paper Radar

Posterior Flow Matching

Chenrui Ma, Xi Xiao, Lin Zhao, Xinru Jiang, Tianyang Wang, Yanning Shen, Ferdinando Fioretto

cs.LG cs.AIcs.CV ·submitted 2026-09-26 20:44 UTC · v1
hot direction

Abstract

Flow matching learns how to transport noise toward data, but its mean prediction does not directly report the endpoint uncertainty remaining at an intermediate state. This paper introduces Posterior Flow Matching (PFM), a general generative modeling framework that learns conditional velocity means and componentwise variances with a Gaussian posterior model. The same model evaluation supplies both a transport prediction and endpoint uncertainty, making posterior diagnostics available throughout generation. These learned moments support a mean-driven ordinary differential equation (ODE) and a posterior-driven stochastic differential equation (SDE), whose local noise scales adapt to the predicted uncertainty. Across image, point-cloud, molecular, and action generation, PFM improves generation quality and provides endpoint-posterior readouts throughout generation.

Closest papers

Nearest neighbours by meaning across the whole index.

Beyond Random Couplings: Contrastive Noise Alignment in Generative Flows

cs.CV cs.LG ·2026-09-16 ·Lennart Wittke, Vinicius Azevedo
similarity 0.68
Abstract · 21 pages, 10 figures, 9 tables

Diffusion and flow-matching models are typically trained by corrupting data through independently sampled Gaussian noise. While simple and scalable, this forward process induces arbitrary data-noise couplings, forcing the network to learn high-curvature transports between unrelated endpoints. Existing optimal-transport methods reduce this burden by reassigning fixed noise samples to data, but the source noise distribution itself remains passive. To address this, we introduce Contrastive Noise Alignment (CNA), a training-time method that creates dynamic, contrastive couplings by optimizing the noise representations directly. By modeling the noise batch as an interacting particle system, CNA employs a cross-modal InfoNCE objective to align noise particles with their paired data targets. To prevent spatial collapse, this alignment is regularized using an angular entropy term and a radial norm penalty. We show theoretically that this equilibrium asymptotically preserves Gaussian structures, maintaining tractability during inference. Empirically, CNA improves the alignment between noise and data, reduces flow curvature, and provides better generation quality with fewer required sampling steps. For few-step, pixel-space generation (2-4 NFEs), CNA reduces FID by over 50\% compared to standard rectified flow, and by at least 24\% against Optimal Transport baselines.

0.17

Empirical Bayes Flow Matching for Continuous Cryo-EM Heterogeneity

cs.LG q-bio.BM ·2026-11-03 ·Daniel Aibinder, Azmi Haider, Dan Rosenbaum
top venue similarity 0.67
Abstract · Accepted at NeurIPS 2026. 20 pages, 6 figures, 5 tables. Code: https://daiyral.github.io/Empirical-Bayes-Flow-Matching-for-Continuous-Cryo-E…

We address the problem of learning generative priors over high-dimensional latent variables from indirect, noisy observations, with a focus on continuous heterogeneity in cryo-electron microscopy (cryo-EM). In this setting, 3D atomic structures $x \in \mathbb{R}^{3N_a}$ are never directly observed and must be inferred from 2D projection images $y$ generated by a complex, non-linear forward operator. We propose Empirical Bayes Flow Matching (EB-FM), a method that learns a flow matching prior $p_θ(x)$ directly from observations using a Monte Carlo stochastic approximation Expectation-Maximization framework. EB-FM alternates between an E-step that performs approximate posterior sampling from $p_θ(x \mid y)$ via guided flow-based generation while maintaining persistent latent estimates through stochastic approximation, and an M-step that updates the flow model using these stabilized latent variables as training targets. This approach eliminates the need for pretraining on clean data and naturally accommodates non-linear forward models. We validate our method on controlled inverse problems using two-moons and MNIST data, and demonstrate its effectiveness on simulated cryo-EM data with known poses, showing that EB-FM can recover continuous distributions of protein conformations directly in atomic-coordinate space.

0.48

Spatially-Grounded Flow Matching: Structured Source Distributions for Image Generation

cs.CV ·2026-08-16 ·Arman Zarei, Mahdi M. Kalayeh
similarity 0.65
Abstract

Current flow matching models learn to transport the source i.i.d. Gaussian noise into the target distribution of natural images, yet this source distribution carries no notion of spatial structure. Images however are fundamentally local since nearby pixels are strongly correlated. By sampling the noise independently, we hypothesize that models are implicitly encouraged to exploit less noisy neighbors as context during training, partially bypassing the need to properly learn the true local structure of images. The source distribution, in other words, works against the inductive bias of the image domain. To ameliorate this design discrepancy, we propose StructFlow which encodes spatial locality directly into the source by having the pixels within a small region share a common noise component. This structured source produces transport paths that are geometrically aligned with image regions - enabling properties that generic flow matching struggles to provide: fine-grained local editing that naturally respects boundaries, robust structure preservation, and smooth semantic interpolation between images. We show that these benefits also extend to large pre-trained models, demonstrating that StructFlow can even be incorporated through a lightweight post-training phase. Comprehensive experiments on multiple datasets, in unconditional, class and text-conditioned regimes, using different diffusion transformer architectures confirm that StructFlow not only offers competitive image generation quality, but also significantly improves localized controllable re-synthesis.

0.34

Beyond Straightness: Non-Crossing Flow Matching via Quantile AlignTree Coupling

cs.LG ·2026-09-03 ·Junyi Lin, Mengyu Li, Jingxuan Hu, Kejun He, Cheng Meng
benchmarktheory similarity 0.65
Abstract

The performance of Flow Matching largely depends on the quality of the coupling between the source and target distributions. However, independent coupling often leads to path crossings and local velocity ambiguity, while OT-based couplings typically incur high construction costs. To address this challenge, we propose Quantile AlignTree Flow Matching (QAT-FM), an efficient structured coupling strategy that constructs a hierarchical coupling between a Gaussian prior and the target data distribution via a quantile-aligned tree structure. QAT-FM constructs the coupling in $\mathcal{O}(Nd\log N)$ time and supports per-pair source sampling with $\mathcal{O}(d)$ complexity, enabling scalable training for large-scale high-dimensional generative tasks. Theoretically, we prove that the QAT coupling satisfies marginal consistency, induces non-crossing linear interpolation paths, and consistently improves path separation at intermediate times compared with independent coupling, thereby alleviating local velocity ambiguity. QAT-FM further extends naturally to conditional generation, enabling structured conditional coupling while preserving global Gaussian alignment. Experiments across diverse benchmark datasets demonstrate that QAT-FM achieves competitive generative performance while substantially reducing coupling construction cost.

0.19

Calibrating Generative Flows through Regional Probability Matching

eess.IV ·2026-10-12 ·Yubo Huang, Xiongyu Hu, Xin Lai, Zixi Wang, Liao Jian +2 more
benchmarkhot direction similarity 0.65
Abstract

Generative models offer a powerful route to scientific simulation and discovery by enabling the study of complex populations and rare phenomena. Flow matching can generate realistic samples, yet inaccurate occurrence frequencies can distort population statistics and compromise inferences about rare events. Here we developed generalized mass-preserving coupled flow matching (G-MPCF), which matches the probability masses of Gaussian source regions to those of weighted data regions. This construction encodes the desired regional frequencies in the correspondence between noise and observations, shaping the learned dynamics while preserving the prescribed target distribution. On a rare-mode benchmark, G-MPCF reduces relative score error by 95.0% and numerical integration error by 89.9% compared with independent flow matching. On CIFAR-10 with a 100:1 class-probability ratio, it reduces corrected class-mass error by 44.6% compared with minibatch optimal transport, while improving mean image fidelity with matched computational budgets. We envision that incorporating regional probability calibration into generative modelling will support more reliable scientific simulation and inference, enabling the study of systems in which rare events and their occurrence frequencies are central to understanding the underlying mechanisms.

0.32

Deterministic Prediction as Global Velocity: Flow Matching for Probabilistic Time-Series Forecast

cs.LG ·2026-10-13 ·Jiajing Hu, Xu Han, Li-Ping Liu
top venuecode/data similarity 0.65
Abstract

Accurate uncertainty quantification is critical for many time series forecasting applications, and recent studies have increasingly adopted generative models to capture predictive uncertainty. Existing approaches often combine deterministic forecasting with generative modeling to improve predictive performance. However, their joint training is typically unstable because the deterministic model is embedded within the generative process, resulting in deep computational graphs and changing the probability transport path. In this work, we propose DV-Flow, a unified flow-matching formulation that incorporates the deterministic predictor as a parallel velocity component rather than as the endpoint of the transport path. This design enables stable end-to-end optimization while subsuming several existing formulations as special cases, allowing direct comparisons among them. Extensive experiments on six real-world datasets demonstrate that DV-Flow consistently improves probabilistic forecasting performance over strong baselines. Our code is available at https://github.com/tufts-ml/dv-flow-time-series.

0.36

Escaping the Curse of Dimensionality in One-Step Flow-Based Generative Models

cs.LG math.ST ·2026-11-11 ·Konstantin Yakovlev
top venue similarity 0.64
Abstract · 52 pages; accepted at NeurIPS 2026

Flow and diffusion-based generative models achieve state-of-the-art quality, but require multiple forward passes at inference time, making them computationally costly. Consistency models overcome this by learning a flow map directly, enabling few-step generation and dramatically reducing the computation needed at test time. However, formal statistical guarantees for such one-step generators are scarce. Existing results either suffer from the curse of dimensionality or rely on impractical loss functions. We investigate whether one-step flow-based models can escape the curse of dimensionality under the assumption that the data distribution is generated by a low-dimensional surface perturbed with Gaussian noise. In this setup, we derive approximation and generalization error bounds for the flow map estimate that circumvent the curse of dimensionality. Our findings provide the first principled explanation for practical success of one-step flow-based generative models.

0.32

Plug-and-Play ADMM for Inverse Problems with Flow Matching Denoiser

cs.LG ·2026-11-10 ·Dibyanshu Kumar, Magda Gregorova
top venue similarity 0.64
Abstract · Accepted at NeurIPS 2026

Plug-and-Play (PnP) methods with generative priors have achieved strong performance in inverse problems, with recent flow matching models enabling fast and scalable reconstruction. However, a key challenge in PnP methods is maintaining consistency between data fidelity updates and the learned generative prior, as data-consistency steps can move iterates away from the flow trajectory. We propose an ADMM-based PnP framework for flow matching priors that enforces stronger consistency between the measurement model and the prior through auxiliary variables and dual updates. In addition, we introduce a trajectory correction mechanism that restores consistency with the flow via a combination of deterministic and stochastic projections. Our method provides a more principled integration of flow priors within PnP frameworks, achieving high-quality reconstructions with reduced computational cost.

0.35