AI Paper Atlas

Research map / Probabilistic and scientific modeling / Variational and generative sampling

Posterior Augmented Flow Matching

George Stoica, Sayak Paul, Matthew Wallingford, Abhay Nori, Vivek Ramanujan, Winson Han, Ali Farhadi, Ranjay Krishna, Judy Hoffman

ECCV 2026 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Flow matching and flow sampling · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Train with PAFM (posterior-weighted multi-target supervision) instead of FMMMDiT · CC12MFID50K: 10.37 -> 9.45Improved
Use PAFM with nearest-neighbor targets instead of FMSiT-B/2 and SiT-XL/2 · ImageNet-1K 256FID, IS, sFID, precision, recall: up to 3.4 FID50KImproved
Vary number of neighbors K for PAFM (ablation)SiT-B/2 · ImageNet-1K 256FID: best at K=16Mixed
Use random-crop augmentation targets (aug scale 1.25) in PAFM (ablation)SiT-B/2 (K=5) · ImageNet-1K 256FID: -3.42 FID at scale 1.25Improved
Sample K times from VAE moment distribution as targets (ablation)SiT-B/2 · ImageNet-1K 256FIDMixed
Use PAFM without REPA loss (ablation)SiT-B/2 · ImageNet-1K 256 (no REPA)FID, sFID, precision, recallImproved
Use PAFM instead of FM for gradient varianceREPA-SiT-B/2 (K=16) · ImageNet-1K 256mean batch-gradient variance: 0.80 -> 0.22Improved
Use PAFM instead of FM with nearest neighbors precomputedMMDiT · CC12M training costthroughput: -6.6% throughput, +0.4% peak memoryHurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy