AI Paper Atlas

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

Streamlined Variational Inference with Higher Level Random Effects

Tui H. Nolan, Marianne Menictas, Matt P. Wand

JMLR 2020 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Variational inference theory and gradients · 6 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 streamlined QR-based message passing instead of naive implementationMean field VB, two-level model · m=1,600 groupsmedian elapsed time (s): 427.00 -> 0.183Improved
use streamlined instead of naive implementationMean field VB · m=200 groupsmedian elapsed time (s): 2.75 -> 0.035Improved
increase number of groups mStreamlined Algorithm 1 · m 400 to 32,400average elapsed time (s): 0.0781 -> 6.980Hurt
use streamlined MFVB instead of Stan MCMCMean field VB · Perinatal datafitting time: 36 hours -> 24 secondsImproved
use streamlined Gibbs sampling in R instead of StanMCMC · Perinatal datafitting time: 36 hours -> 3.5 hoursImproved
implement streamlined Gibbs in C++ rather than RMCMC · Perinatal datafitting time: 3.5 hours -> 4.9 minutesImproved
use streamlined mean field VB instead of MCMCMean field VB · Perinatal datacredible intervalsNo clear effect
apply mean field product restrictionStreamlined mean field VB · Simulated two-level data95% credible interval coverage: 93.5 to 96.2No 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