AI Paper Atlas

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

Double Spike Dirichlet Priors for Structured Weighting

Huiming Lin, Meng Li

JMLR 2022 · Probabilistic and scientific modeling › Uncertainty and Gaussian processes · approach: Nonparametric and PAC-Bayes Bayesian learning · 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
use double spike Dirichlet prior instead of symmetric DirichletBayesian structured weighting · Simulation Scenario 1l1 error: 0.428 -> 0.032Improved
use joint double spike Dirichlet instead of two-step lassoStructured simplex estimation · Simulation Scenario 1l1 error: 0.139 -> 0.032Improved
use double spike Dirichlet instead of two-step lassoStructured simplex estimation · Simulation Scenario 2l1 error: 0.298 -> 0.210Improved
use double spike Dirichlet instead of symmetric DirichletStructured simplex estimation · Simulation Scenario 2l1 error: 0.436 -> 0.210Improved
larger rho1 and smaller rho2 (ablation)Double spike Dirichlet · Scenario 1 sensitivityl1 error: l1 error decreases monotonically, indistinguishable for large rho1Improved
larger rho1 and smaller rho2 (ablation)Double spike Dirichlet · Scenario 2 sensitivityl1 error: monotone decreasing pattern no longer holdsNo clear effect
fixed theta=1/K instead of Beta prior on theta (ablation)Double spike Dirichlet · Simulation, deviation levelsl1 error: 0.420 vs 0.432 at deviation 0.3155Improved
two-step horseshoe instead of two-step lassoTwo-step baseline · Simulation, deviation levelsl1 error: 0.464 -> 0.453 at deviation 0.3155Improved
increase signal-to-noise ratio from 0.6 to 2.0Double spike Dirichlet · Simulation SNRl1 error: 0.366 -> 0.017Improved
double spike Dirichlet instead of simple averageForecast combination · ECB SPF GDP forecastout-of-sample RMSE: 1.502 -> 1.441Improved
double spike Dirichlet instead of two-step lassoForecast combination · ECB SPF GDP forecastout-of-sample RMSE: 1.542 -> 1.441Improved
random theta instead of fixed theta (ablation)Double spike Dirichlet · ECB SPF GDP forecastout-of-sample RMSE: 1.441 vs 1.444No 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