AI Paper Atlas

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

On two ways to use determinantal point processes for Monte Carlo integration

Guillaume Gautier, Rémi Bardenet, Michal Valko

NeurIPS 2019 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Importance sampling and quadrature

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 EZ estimator instead of BH estimatorDPP Monte Carlo integration, bump function · d <= 2variance decayImproved
use EZ estimator instead of BH estimatorDPP Monte Carlo integration, bump function · d >= 3variance decayHurt
increase number of DPP points N to reach number of eigenfunctions M in the integrandEZ estimator · sum of eigenfunctions, M=70, d=1 to 4varianceImproved
increase N with M=N+1 so the integrand is never fully in the kernel spanEZ estimator · sum of eigenfunctions, M=N+1variance decay: order 1/N^2Improved
apply EZ estimator to non-smooth or discontinuous functions in higher dimensionEZ estimator · discontinuous functions, Appendix B.3-B.6varianceHurt
compare EZ against BH normality of estimator (Kolmogorov-Smirnov test)EZ estimator · bump function, N=300KS p-value: 0.03 to 0.24 vs 0.60 to 0.99Hurt
use vanilla Monte Carlo instead of DPP estimators when integrand is non-smoothDPP estimators · non-smooth functions, small NvarianceMixed

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