流匹配中的流模型究竟带来了什么？
What Does a Stream Model Buy You in Flow Matching?
- RIKEN（理化学研究所）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
本文通过理论约简、实验分析和代码审计，揭示流级流匹配中的GP流模型并未带来实质收益，其设计空间坍缩为两条标量曲线，且发布的代码未实现所述机制。
AI中文摘要：
流级流匹配用高斯过程（GP）流连接每个源-目标对，替代条件流匹配（CFM）的线性插值，并在2-Gaussian、MNIST和CIFAR-10基准上报告了比ICFM更低的样本误差。我们探究这种流模型实际贡献了什么。三个结果回答了这个问题。(i) 约简。流级CFM目标仅通过$(s_t,\dot{s}_t)$的逐时联合分布依赖于流定律，因此高斯流能达到的条件路径恰好是CFM已经参数化的高斯条件路径；在gpcfm实际使用的坐标式、共享标量核构造中，整个设计空间坍缩为两条标量曲线$(m_t,v_t)$，跨时间协方差仅影响估计器方差。(ii) GP是该空间的受限图。一个核同时设定$m_t$和$v_t$，因此论文自己提出的扩大覆盖范围的配方——收缩SE长度尺度——破坏了插值（中点均值权重从1.03降至0.00）。在2-Gaussian基准上，这使得GP图在高覆盖度下在15/200次运行中发散，而解耦的$(m_t,v_t)$图为0/200（$p=6.6\times10^{-5}$），且两条曲线的交叉表明发散跟踪均值而非方差。在MNIST上，同样的扫描不发散且顺序反转，因此耦合是否有害取决于基准；两者共同的是该配方毫无收益——没有覆盖水平超过论文自身的，且超过$\max_t\sqrt{v_t}\approx0.6$后两种图都退化。(iii) 审计。发布的代码未实现其描述的机制：状态和速度独立采样（相关性$0.00\pm0.01$，而预期为$\pm0.83$–$0.99$）。
英文摘要：
Stream-level flow matching replaces the linear interpolant of conditional flow matching (CFM) by a Gaussian-process (GP) stream connecting each source--target pair, and reports lower sample error than \icfm{} on 2-Gaussian, MNIST and CIFAR-10 benchmarks. We ask what such a stream model actually contributes. Three results answer the question. (i)~\emph{Reduction.} The stream-level CFM objective depends on the stream law only through the per-time joint law of $(s_t,\sdot_t)$, so the conditional paths a Gaussian stream can reach are exactly the Gaussian conditional paths CFM already parametrises; in the coordinate-wise, shared-scalar-kernel construction gpcfm actually uses, the entire design space collapses to two scalar curves $(m_t,v_t)$, and cross-time covariance affects only estimator variance. (ii)~\emph{The GP is a constrained chart of that space.} One kernel sets both $m_t$ and $v_t$, so the paper's own recipe for widening coverage-shrinking the SE length-scale---destroys the interpolant (the midpoint mean weight falls from $1.03$ to $0.00$). On the 2-Gaussian benchmark this makes the GP chart diverge on $15/200$ runs at high coverage against $0/200$ for a decoupled $(m_t,v_t)$ chart ($p=6.6\times10^{-5}$), and crossing the two curves shows the divergence tracks the mean, not the variance. On MNIST the same sweep does not diverge and the ordering reverses, so whether the coupling is harmful is benchmark-dependent; what holds on both is that the recipe buys nothing---no coverage level beats the paper's own, and past $\max_t\sqrt{v_t}\approx0.6$ both charts degrade. (iii)~\emph{Audit.} The released code does not implement the mechanism it describes: state and velocity are drawn independently ($\mathrm{corr}=0.00\pm0.01$ against an intended $\pm0.83$--$0.99$).
