从方差视角重新审视流匹配，提出稳定速度新方法，提升生成模型训练稳定性和性能。
arXiv:2602.05435v2 Announce Type: replace Abstract: While flow matching is elegant, its reliance on single-sample conditional velocities leads to high…
Flow Matching结合优化子类先验，为医学图像增强提供新方案。
arXiv:2605.16469v1 Announce Type: cross Abstract: Rare diseases dominate the diagnostic challenge in medical imaging yet are severely underrepresented…
流匹配新方法学习无偏排列，排序与图匹配任务或迎新突破。
arXiv:2605.16755v1 Announce Type: new Abstract: Learning permutations is fundamental to sorting, ranking, and matching, but existing differentiable me…
联合人脸运动与语音生成，Flow Matching实现音视频同步合成，突破传统分离建模局限。
arXiv:2506.23552v2 Announce Type: replace Abstract: The intrinsic link between facial motion and speech is often overlooked in generative modeling, wh…
提出Flow-OPD新方法，用同策略蒸馏解决流匹配模型在多任务对齐中的奖励稀疏和梯度干扰问题。
arXiv:2605.08063v3 Announce Type: replace-cross Abstract: Existing Flow Matching (FM) text-to-image models suffer from two critical bottlenecks under …
揭示WarmPrior用动作历史替代高斯噪声，大幅提升机器人操作成功率，简单有效。
arXiv:2605.13959v1 Announce Type: cross Abstract: Generative policies based on diffusion and flow matching have become a dominant paradigm for visuomo…