多篇研究论文探讨了生成模型中流匹配技术的进展。一项研究引入了“最小时间梯度流”方法来优化风险降低速度,提出了一种独特的摆线形最小化器。另一篇论文修正了高斯流匹配的条件最优传输(CondOT)调度,展示了改进的采样精度和收敛界限。此外,研究提出了一种用于文本到语音合成的条件流匹配的频率选择性增强策略,增强了频谱相干性并降低了计算成本。其他工作将各种神经表征不相似性度量统一在流匹配框架下,并提出了变分终端速度流匹配(VTV-FM),以改进二阶动力学中的传输几何和生成质量。
AI
arXiv:2610.01426v1 Announce Type: new Abstract: Prescribing the speed of gradient flow on the risk itself, by the dynamics $\dot w=-u(E(w))\nabla E(w)/\abs{\nabla E(w)}^{2}$, makes the risk $e(t)=E(w(t))$ obey $\dot e=-u(e)$ exactly, whatever the landscape~$E$; the time needed to…
arXiv cs.LG
TIER_1English(EN)·Ron Levy, Michael Elad·
arXiv:2609.39488v1 Announce Type: new Abstract: Flow matching generates samples by gradually transforming noise into data. In practice, using a finite number of sampling steps introduces a numerical error that depends on the chosen schedule. We study this dependence for Gaussian …
arXiv cs.LG
TIER_1English(EN)·Isha Pandey, Varad Deshpande, Abhijat Bharadwaj, Ganesh Ramakrishnan·
arXiv:2609.34431v2 Announce Type: replace Abstract: Conditional Flow Matching (CFM) models for text-to-speech (TTS) suffer from incoherent frequency evolution during inference. While similar spectral imbalances are addressed in diffusion models for other domains, those generic so…
arXiv cs.AI
TIER_1English(EN)·Zeyuan Ye, Xue-Xin Wei·
arXiv:2609.31544v1 Announce Type: new Abstract: Neural representational dissimilarity quantifies differences between neural response distributions, and is essential for comparing neural codes across stimuli, brain areas, tasks, and models. Commonly used distance metrics involve d…
arXiv cs.CV
TIER_1English(EN)·Haoyang Jiang, Yuheng Li, Di Yang, Yanhai Xiong, Haipeng Chen, Yi He·
arXiv:2610.00785v1 Announce Type: new Abstract: Flow matching (FM) learns generative transport by fitting continuous-time motion from a simple source distribution to the data distribution. Most existing methods use first-order bridges: once a source and a target sample are paired…
arXiv cs.CV
TIER_1English(EN)·Ziqi Jiang, Zhenqi He, Long Chen·
arXiv:2610.01408v1 Announce Type: new Abstract: Trajectory crossing remains a critical bottleneck in Flow Matching (FM), and previous works typically view these crossings from a theoretical optimization perspective causing velocity averaging. They attempt to address it indirectly…
arXiv:2609.38918v1 Announce Type: cross Abstract: Flow Matching (FM) learns a velocity field whose ODE transports a simple source distribution to a target law. Existing finite-sample theory largely treats ambient-space regularity or data supported exactly on low-dimensional sets.…