We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat.ML

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo

Statistics > Machine Learning

Title: Tree pyramidal adaptive importance sampling

Abstract: This paper introduces Tree-Pyramidal Adaptive Importance Sampling (TP-AIS), a novel iterated sampling method that outperforms state-of-the-art approaches like deterministic mixture population Monte Carlo (DM-PMC), mixture population Monte Carlo (M-PMC) and layered adaptive importance sampling (LAIS). TP-AIS iteratively builds a proposal distribution parameterized by a tree pyramid, where each tree leaf spans a subspace that represents its importance density. After each new sample operation, a set of tree leaves are subdivided for improving the approximation of the proposal distribution to the target density. Unlike the rest of the methods in the literature, TP-AIS is parameter free and requires no tuning to achieve its best performance. We evaluate TP-AIS with different complexity randomized target probability density functions (PDF) and also analyze its application to different dimensions. The results are compared to state-of-the-art iterative importance sampling approaches and other baseline MCMC approaches using Normalized Effective Sample Size (N-ESS), Jensen-Shannon Divergence, and time complexity.
Comments: 20 pages + 13 pages of additional result plots and evaluation details
Subjects: Machine Learning (stat.ML); Machine Learning (cs.LG); Computation (stat.CO)
Cite as: arXiv:1912.08434 [stat.ML]
  (or arXiv:1912.08434v2 [stat.ML] for this version)

Submission history

From: Javier Felip Leon [view email]
[v1] Wed, 18 Dec 2019 08:05:07 GMT (6782kb,D)
[v2] Mon, 23 Mar 2020 18:55:24 GMT (7322kb,D)

Link back to: arXiv, form interface, contact.