Flow matching research advances generative modeling and inverse problems · 10 sources tracked
ByPulseAugur Editorial·[17 sources]·
Recent research explores advancements in flow matching techniques for generative modeling and inverse problems. Papers introduce FUSE for efficient multimodal simulation-based posterior estimation, Diagonal Flow Matching (Diag-CFM) for stable inverse design with uncertainty quantification, and Lagrangian Dual Flows for constrained generation. Other work focuses on score-regularized joint sampling for improved expectation estimation and asymptotic-preserving analysis of diffusion and flow-matching samplers. Additionally, flow matching is being applied to sparse-view CT reconstruction and geophysical inversion, demonstrating its versatility across various scientific and engineering domains.
AI
IMPACT
Advances in flow matching techniques are enhancing generative modeling capabilities and enabling more efficient solutions for complex inverse problems across scientific and engineering fields.
RANK_REASON
Multiple arXiv papers detailing novel research in flow matching techniques and their applications.
arXiv:2607.05252v1 Announce Type: new Abstract: Simulation-Based Inference (SBI) is critical for scientific discovery, with generative models offering a promising path toward efficient inference. However, existing methods struggle with effective multimodal modeling. They often re…
arXiv cs.LG
TIER_1English(EN)·Miguel de Campos, Werner Krebs, Hanno Gottschalk·
arXiv:2607.04513v1 Announce Type: cross Abstract: Flow matching is a powerful tool for generative modeling, but emerging applications in robotics, planning, and physics require inference-time constraints on generated outputs. Such constraints are often complex and highly nonlinea…
arXiv:2607.04113v1 Announce Type: new Abstract: Diffusion and flow-matching samplers integrate a learned probability-flow ODE from a large noise scale down to a small terminal floor $\sigma_{\min}$, at which the score is stiff and the flow develops a boundary layer. We treat $\si…
arXiv cs.AI
TIER_1English(EN)·Jiayang Shi, Lincen Yang, Zhong Li, Tristan van Leeuwen, Daniel M. Pelt, K. Joost Batenburg·
Simulation-Based Inference (SBI) is critical for scientific discovery, with generative models offering a promising path toward efficient inference. However, existing methods struggle with effective multimodal modeling. They often rely on brute-force fusion strategies that ignore …
MrFlow accelerates text-to-image diffusion by combining low-resolution generation with pixel-space super-resolution and noise injection, achieving up to 25x speedup without training or runtime modifications.
arXiv cs.LG
TIER_1English(EN)·Baldur Paulwitz, Stefan Buske·
arXiv:2606.31288v1 Announce Type: new Abstract: We demonstrate the application of Flow Matching, a technique originating from generative Artificial Intelligence, to probabilistic inversion in geophysical settings, such as seismic Full-Waveform inversion. We adapt the well-establi…
We demonstrate the application of Flow Matching, a technique originating from generative Artificial Intelligence, to probabilistic inversion in geophysical settings, such as seismic Full-Waveform inversion. We adapt the well-established mathematical theory of Flow Matching from g…
arXiv cs.AI
TIER_1English(EN)·Francois Porcher, Nicolas Carion, Karteek Alahari, Shizhe Chen·
arXiv:2606.29059v1 Announce Type: cross Abstract: World modeling requires forecasting uncertain futures while preserving information useful for downstream perception. Existing visual world models often struggle to satisfy both goals: VAE-based stochastic models operate in low-dim…
arXiv cs.LG
TIER_1English(EN)·Liam A. Kruse, Houjun Liu, Alexandros E. Tzikas, Mansur M. Arief, Mykel J. Kochenderfer·
arXiv:2606.29724v1 Announce Type: new Abstract: Normalizing flows are powerful generative models that learn an invertible mapping between complex data distributions and simple latent distributions, typically a standard normal density. However, this choice of latent density can im…
Normalizing flows are powerful generative models that learn an invertible mapping between complex data distributions and simple latent distributions, typically a standard normal density. However, this choice of latent density can impose unnecessary complexity on the learned flow …
arXiv:2607.03524v1 Announce Type: new Abstract: We propose Perceptual Flow Matching (PFM), a simple yet effective framework for few-step generation in flow-matching models. Rather than performing velocity regression in the conventional VAE latent space, PFM supervises flow matchi…
arXiv:2607.01642v1 Announce Type: new Abstract: Hardware-agnostic strategies for accelerating text-to-image diffusion, such as timestep distillation and feature caching, can reduce inference time without custom kernels or system-level optimization. Among them, multi-resolution ge…
