AI Paper Atlas

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

The Boomerang Sampler

Joris Bierkens, Sebastiano Grazzi, Kengo Kamatani, Gareth Roberts

ICML 2020 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Hamiltonian Monte Carlo · 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 control-variate subsampling instead of full-data gradientsBoomerang Sampler · Logistic regression, large nESS per secondImproved
Use control-variate subsampling instead of the naive unbiased gradient estimatorSubsampled Boomerang Sampler · Large-data scalingevent rateImproved
Use factorised component-wise switchingBoomerang Sampler · Sparse conditional-dependence modelscomputational costImproved
Use Boomerang dynamics rather than constant-velocity dynamicsPDMC samplers · Logistic regressionESS per secondImproved
Increase dimension while using subsampled algorithmsSubsampled Boomerang Sampler · Logistic regressionESS per secondHurt
Use Boomerang rather than MALALogistic regression samplers · Dimensions d ≤ 32ESS per second: d ≤ 32Improved
Increase dimension when comparing Boomerang with MALABoomerang Sampler and MALA · Logistic regressionscaling with dimensionHurt
Use the Boomerang Sampler instead of BPSBoomerang Sampler · Standard-normal target with perturbed reference measureESS per secondImproved
Perturb the reference mean or covariance away from the target GaussianBoomerang Sampler · Standard-normal targetESS per secondMixed
Use Boomerang rather than Zig-Zag for high-level diffusion-bridge coefficientsFactorised Boomerang Sampler · Diffusion bridges, N=10number of reflections: N=10Improved
Increase the number of observations to n ≥ 10,000Boomerang Sampler · Logistic regression, d=2accepted switches: n ≥ 10,000; d = 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