Search papers, labs, and topics across Lattice.
This paper introduces co-evolving path and flow networks trained via a path-flow alignment objective for flow matching, where the flow learns path velocity and the path aligns to flow velocity. The authors identify path overfitting, characterized by low-entropy bottlenecks in induced probability paths, as a failure mode. A novel stochastic path regularizer is proposed to mitigate this by introducing an entropy floor for path marginals, leading to improved FID scores on ImageNet-256x256 with SiT backbones.
Jointly training path and flow networks with a novel alignment objective unlocks improved generative model performance by preventing path overfitting and low-entropy bottlenecks.
We study path-flow alignment as a unified training objective for flow matching. Instead of fixing the interpolation path and learning only the velocity field, we jointly train an endpoint-preserving path network and a flow network using the same alignment loss: the flow learns to match the path velocity, and the path learns to align its velocity to the current flow. Although every fixed learned path defines a valid flow-matching objective, the alignment loss alone is not a reliable criterion for path learning. We identify path overfitting, a failure mode in which the alignment loss decreases while sample quality worsens. We find that this failure is associated with low-entropy bottlenecks in the induced probability path, where the learned path routes samples through overly concentrated intermediate marginals. Motivated by this diagnosis, we introduce a stochastic path regularizer that hides part of the source information from the path network while preserving exact endpoints. The resulting regularization gives an explicit entropy floor for the stochastic training-path marginals and empirically suppresses the bottleneck in the learned sampler, making joint path-flow training effective. On ImageNet-256x256 with SiT backbones, our method consistently improves FID across model scales, extends to model-guidance training, and leaves the inference-time architecture and sampler unchanged. Code is available at https://github.com/lizeyu090312/traj_opt_paper
