AI Paper Atlas

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

Learning Optimal Flows for Non-Equilibrium Importance Sampling

Yu Cao, Eric Vanden-Eijnden

NeurIPS 2022 · 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 flow velocity by minimizing estimator variance with assisted SGDNEIS · 2D asymmetric 2-mode Gaussian mixturevariance of Z1 estimator: 10^6 -> 1Improved
train flow velocity by minimizing estimator variance (gradient ansatz, 60 SGD steps)NEIS · 10D symmetric 4-mode Gaussian mixturevariance of Z1 estimator: 2.15\times 10^6 -> 10Improved
NEIS with optimized flow instead of AIS-100 under equal query budgetNEIS · 2D Gaussian mixturestd of Z1 estimateImproved
NEIS with optimized flow instead of AIS-100 under equal query budgetNEIS · 10D Gaussian mixturestd of Z1 estimateImproved
use two-parametric ansatz instead of generic linear ansatz for velocity field (ablation)NEIS · 10D funnel distributionZ1 estimation accuracyImproved
use generic linear ansatz NEIS instead of AIS-100NEIS · 10D funnel distributionZ1 estimation accuracyImproved
use asymmetric t_-=0 instead of symmetric t_-=-1/2 (ablation)NEIS · 10D funnel distributionZ1 estimation accuracyHurt
bias from ODE time discretizationNEIS · Gaussian mixturesbias of Z1 estimateNo clear effect

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