AI Paper Atlas

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

Nonparametric learning from Bayesian models with randomized objective functions

Simon Lyddon, Stephen Walker, Chris C Holmes

NeurIPS 2018 · Probabilistic and scientific modeling › Uncertainty and Gaussian processes · approach: Nonparametric and PAC-Bayes Bayesian learning · 29 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
Correct the mean-field VB posterior using the posterior bootstrapVB logistic regression · German Creditposterior uncertainty calibrationImproved
Use VB-NPL rather than MCMC samplingBayesian logistic regression · German Creditruntime for 1 million samples: 33 minutes -> 21 seconds; 95 timesImproved
Increase c to give more weight to external prior pseudo-samplesBRF · 13 UCI binary classification datasetstest accuracyImproved
Set c around the number of external training samplesBRF · 13 UCI binary classification datasetstest accuracyImproved
Set c=0, removing the model and prior pseudo-samples (ablation)BRF · Bank datasetmean classification accuracy: 0.895No clear effect
Decrease c to reduce reliance on the centering modelVB-NPL · Bivariate normal posterior examplecovariance structure recoveryImproved
Use KL objective d_KL(p,q) instead of d_KL(q,p) (ablation)VB-NPL · Bivariate normal posterior exampleposterior uncertainty correctionMixed

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