AI Paper Atlas

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

Pathfinder: Parallel quasi-Newton variational inference

Lu Zhang, Bob Carpenter, Andrew Gelman, Aki Vehtari

JMLR 2022 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Variational inference theory and gradients · 25 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
replace ADVI (mean-field or dense) with single-path Pathfinderposterior approximation · posteriordb (20 models)1-Wasserstein distance: median distance of mean-field ADVI more than double for 8 of 20 models; dense ADVI 9 of 20Improved
use Pathfinder instead of Stan phase I warmupposterior approximation · posteriordb (20 models)1-Wasserstein distance: 7 of 20 models with more than double the distanceImproved
use Pathfinder instead of Stan phase I and ADVIcomputational cost · posteriordblog density and gradient evaluations: gradient evaluations 34, 48, 54 times more than PathfinderImproved
use multi-path Pathfinder instead of single-pathPathfinder · bball_drive_event_0-hmm_drive_01-Wasserstein distanceImproved
reduce L-BFGS iterations to 100 and loosen tolerance to 10^-3 (ablation)Pathfinder · posteriordb1-Wasserstein distance: increased more than a factor of 8 for three modelsHurt
increase ELBO Monte Carlo draws from K=5 to K=30 (ablation)Pathfinder · posteriordb1-Wasserstein distance: 3.2% average, 15.1% maximum reduction; 4.9 times log-density evaluationsImproved
increase L-BFGS history from J=6 to J=60 (ablation)Pathfinder · posteriordb1-Wasserstein distance: 0.7% average reduction for 19 models; maximum reduction 16.3%, maximum increase 9.8%Mixed
decrease number of single-path runs from I=20 to I=5 (ablation)multi-path Pathfinder · posteriordb1-Wasserstein distance: median increases 5.4% on average, 0.1% minimum, 12.5% maximumHurt
increase number of single-path runs from I=20 to I=40 (ablation)multi-path Pathfinder · posteriordb1-Wasserstein distance: reduced by 1.1% on average, maximum reduction 5.5%, maximum increase 6.4%No clear effect
use more diffuse initialization uniform(-15,15) instead of uniform(-2,2) (ablation)multi-path Pathfinder · eight schools centered / 100-d Neal funnelapproximate draw coverageImproved
use ADVI instead of Pathfinderposterior approximation · eight schools centered1-Wasserstein distance and ELBOImproved

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