Pith. sign in

REVIEW 3 cited by

On the Asymptotics of Importance Weighted Variational Inference

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2501.08477 v1 pith:XZEESIZ6 submitted 2025-01-14 math.ST stat.TH

On the Asymptotics of Importance Weighted Variational Inference

classification math.ST stat.TH
keywords importancecarlodataexpectationinferencelatentlikelihoodmonte
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

For complex latent variable models, the likelihood function is not available in closed form. In this context, a popular method to perform parameter estimation is Importance Weighted Variational Inference. It essentially maximizes the expectation of the logarithm of an importance sampling estimate of the likelihood with respect to both the latent variable model parameters and the importance distribution parameters, the expectation being itself with respect to the importance samples. Despite its great empirical success in machine learning, a theoretical analysis of the limit properties of the resulting estimates is still lacking. We fill this gap by establishing consistency when both the Monte Carlo and the observed data sample sizes go to infinity simultaneously. We also establish asymptotic normality and efficiency under additional conditions relating the rate of growth between the Monte Carlo and the observed data samples sizes. We distinguish several regimes related to the smoothness of the importance ratio.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. The Wasserstein cost of Importance Sampling

    math.PR 2026-05 unverdicted novelty 7.0

    The expected p-Wasserstein cost of importance sampling is of order n^{-p/d} with matching constants, and the asymptotically optimal proposal is proportional to g^{d/(p+d)}.

  2. Concentration and Calibration in Predictive Bayesian Inference

    stat.ME 2026-05 unverdicted novelty 6.0

    Predictive Bayesian inference posteriors concentrate onto a forward-model-dependent quantity and produce miscalibrated credible sets unless the predictive model contains the true data-generating process.

  3. Predictively Oriented Posteriors

    stat.ME 2025-10 conditional novelty 6.0

    A new posterior family scores the posterior predictive directly, giving slower concentration but better predictive performance under model misspecification.