AI Paper Atlas

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

Meta Flow Maps enable scalable reward alignment

Peter Potaptchik, Adhi Saravanan, Abbas Mammadov, Alvaro Prat, Michael Samuel Albergo, Yee Whye Teh

ICML 2026 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Flow matching and flow sampling · 0 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
use MFM one-step posterior samples instead of GLASS ODE rollouts for posterior samplingMFM-XL/2 · ImageNet 256x256 posterior FID and value-function correlationposterior FID / correlationImproved
steer with MFM-G (N=1) instead of Best-of-N with N=1000MFM-XL/2 class tabby cat · ImageNet steering with ImageReward, PickScore, HPSv2average reward: over 100x fewer NFEsImproved
use gradient-based MFM-G instead of gradient-free MFM-GF (ablation)MFM steering · MNIST class-ratio matchingL2 to target class ratiosImproved
use MFM-Search instead of MFM-GF for gradient-free steering (ablation)MFM steering · ImageNet tabby cat steeringrewardImproved
replace DPS with MFM-based steeringsteering · GMM inverse problem and MNISTdistribution error / class-ratio L2Improved
increase number of Monte Carlo samples in MFM steering (ablation)MFM-G / MFM-GF · GMM inverse problemMMD and Sliced-WassersteinImproved
steer with HPSv2 or PickScore and evaluate on the other reward modelsMFM-G · ImageNet steering, cross-reward evaluationother reward scoresImproved
fine-tune with MFM-FT on HPSv2MFM-XL/2 · ImageNet all classes, HPSv2 prompt templateHPSv2, PickScore, ImageRewardImproved
increase K-step refinement from 1 to 4 steps (ablation)MFM-XL/2 · ImageNet 256x256FID: 3.72 -> 1.97Improved

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