English(EN)Exploring the Design Space of Reward Backpropagation for Flow Matching
新研究推进生成式AI的流匹配模型
作者PulseAugur 编辑部·[14 个来源]·
研究人员正在探索流匹配模型（一种生成模型）的高级技术。一篇论文介绍了渐进式微调（GFT），这是一个基于退火的框架，用于提高将这些模型适应新数据分布时的稳定性和效率。另一种方法LieFlow将对称性发现重新构建为李群上的分布学习问题，从而能够识别连续和离散的对称性。此外，Latent-CFM通过利用预训练的潜在变量模型来提高流匹配效率，尤其适用于高维数据。进一步的理论工作为KL散度和Wasserstein距离下的扩散流匹配（DFM）提供了改进的收敛保证，而拓扑流匹配则将该框架推广到处理具有丰富拓扑结构的有结构数据。
AI
arXiv:2601.22107v2 Announce Type: replace Abstract: We introduce \textit{Prior-Informed Flow Matching (PIFM)}, a conditional flow model for graph reconstruction. Reconstructing graphs from partial observations remains a key challenge; classical embedding methods often lack global…
arXiv cs.LG
TIER_1English(EN)·Nicolas Beltran-Velez, Felix Friedrich, Zhang Xiaofeng, Reyhane Askari-Hemmat, Xiaochuang Han, Adriana Romero-Soriano, Michal Drozdzal·
arXiv:2606.19162v1 Announce Type: new Abstract: Score- and flow-matching models often rely on preference-based reinforcement learning for two purposes: aligning with subjective preferences and, surprisingly, recovering properties such as visual realism and coherent object structu…
We propose SpectralDiT, a lightweight modification to flow-matching Diffusion Transformers that adds timestep-conditioned spectral correction to the MLP residual branch. The module decomposes each residual update into low- and high-frequency components on the patch-token grid, th…
arXiv:2601.22495v2 Announce Type: replace Abstract: Fine-tuning flow matching models is a central challenge in settings with limited data, evolving distributions, or computational constraints. While recent work has produced significant advances, particularly in the area of reward…
Discriminator-Guided Reinforcement Learning (DRL) addresses alignment issues in score- and flow-matching models by using a pretrained representation space discriminator as an optimal reward signal, improving both visual fidelity and semantic quality without human preferences.
arXiv cs.AI
TIER_1English(EN)·Kacper Wyrwal, \.Ismail \.Ilkan Ceylan, Alexander Tong·
arXiv:2606.15897v1 Announce Type: cross Abstract: 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 po…
arXiv cs.AI
TIER_1English(EN)·Yuxuan Chen, Jung Yeon Park, Floor Eijkelboom, Jianke Yang, Jan-Willem van de Meent, Lawson L. S. Wong, Robin Walters·
arXiv:2512.20043v3 Announce Type: replace Abstract: Symmetry is fundamental to understanding physical systems and can improve performance and sample efficiency in machine learning. Both pursuits require knowledge of the underlying symmetries in data, yet discovering these symmetr…
arXiv cs.AI
TIER_1English(EN)·Anirban Samaddar, Yixuan Sun, Viktor Nilsson, Sandeep Madireddy·
arXiv:2505.04486v4 Announce Type: replace-cross Abstract: Flow matching models have shown great potential in image generation tasks among probabilistic generative models. However, most flow matching models in the literature do not explicitly utilize the underlying clustering stru…
arXiv cs.LG
TIER_1English(EN)·Marta Gentiloni Silveri, Giovanni Conforti, Alain Durmus·
arXiv:2606.16610v1 Announce Type: cross Abstract: Diffusion Flow Matching (DFM) has recently emerged as a versatile framework for generative modeling, yet its theoretical convergence properties remain only partially understood. In this work, we provide refined and novel convergen…
arXiv stat.ML
TIER_1English(EN)·Taos Transue, Bohan Chen, So Takao, Bao Wang·
arXiv:2508.13313v4 Announce Type: replace Abstract: Data assimilation (DA) estimates a dynamical system's state from noisy observations. Recent generative models like the ensemble score filter (EnSF) improve DA in high-dimensional nonlinear settings but are computationally expens…
Score- and flow-matching models often rely on preference-based reinforcement learning for two purposes: aligning with subjective preferences and, surprisingly, recovering properties such as visual realism and coherent object structure that matching-based training is intended to l…
We propose SpectralDiT, a lightweight modification to flow-matching Diffusion Transformers that adds timestep-conditioned spectral correction to the MLP residual branch. The module decomposes each residual update into low- and high-frequency components on the patch-token grid, th…
Diffusion Flow Matching (DFM) has recently emerged as a versatile framework for generative modeling, yet its theoretical convergence properties remain only partially understood. In this work, we provide refined and novel convergence guarantees for Brownian motion based DFMs, focu…
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 Euclidean space, overlooking the rich topo…
