AI Paper Atlas

Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes

Bayesian leave-one-out cross-validation for large data

Måns Magnusson, Michael Andersen, Johan Jonasson, Aki Vehtari

ICML 2019 · Probabilistic and scientific modeling › Uncertainty and Gaussian processes · approach: Simulation-based inference · 15 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 simple random sampling with PPS proportional to -log predictive densityLOO estimation on LR(c) 100D, m=10 · LR(c) 100DSE(elpd_loo estimate): 2197.8 -> 11.8Improved
Replace simple random sampling with PPS proportional to -log predictive densityLOO estimation on LR(c) 100D, m=100 · LR(c) 100DSE(elpd_loo estimate): 598.8 -> 4.3Improved
Replace simple random sampling with PPS proportional to -log predictive densityLOO estimation on LR(c) 10D, m=10 · LR(c) 10DSE(elpd_loo estimate): 3310.2 -> 3.2Improved
Use point-estimate likelihood rather than full-posterior predictive density to set PPS probabilities (ablation)PPS LOO estimation · LR(c) 100D, m=100SE(elpd_loo estimate): 4.3 -> 4.5No clear effect
Use point-estimate likelihood rather than full-posterior predictive density to set PPS probabilities (ablation)PPS LOO estimation · LR(c) 100D, m=1000SE(elpd_loo estimate): 1.5 -> 1.5No clear effect
Use point-estimate likelihood rather than full-posterior predictive density to set PPS probabilities (ablation)PPS LOO estimation · LR(c) 10D, m=100SE(elpd_loo estimate): 1.2 -> 1.1No clear effect
Use mean-field variational inference for a correlated posterior (ablation)Bayesian linear regression · LR(c) 100Delpd_loo: -14247 -> -14267No clear effect
Use full-rank ADVI at high parameter dimension (ablation)Bayesian linear regression · LR(i) 100Delpd_loo: -14239 -> -14193Mixed
Increase PPS subsample size from 10 to 100 (ablation)PPS(1) LOO estimation · LR(c) 100DSE(elpd_loo estimate): 11.8 -> 4.3Improved
Increase PPS subsample size from 100 to 1000 (ablation)PPS(1) LOO estimation · LR(c) 100DSE(elpd_loo estimate): 4.3 -> 1.5Improved
Use ADVI(MF) rather than MCMC for Radon model 6 (ablation)Radon model 6 with PPS subsample m=500 · Radonelpd_loo estimate: -16835 -> -16656Mixed
Use PPS sampling instead of simple random sampling as n growsLOO estimation with point-estimate likelihood probabilities · LR(c) 100D, m=10, n=100000SE(elpd_loo estimate): 19721 -> 17.5Improved

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