Invariant Measures as Estimators: Second-Order Stochastic Expansions and Bias Reduction
Abstract
We develop a general second-order asymptotic framework for estimators constructed from invariant probability measures of data-dependent Markov processes. The framework yields a universal bias correction for smooth functionals, computed from the same invariant measure used to construct the estimator, without the resampling and repeated estimation involved in bias-correction methods such as the bootstrap and jackknife. The statistical input is represented by a random one-form, allowing differentials of log-likelihoods and general random criteria, as well as estimating functions, to be treated in a common framework. Under local regularity and localization conditions, we derive second-order stochastic expansions for averages with respect to these invariant measures and for decision rules defined through general loss functions. The expansions separate the contributions of the statistical input, temperature, drift, and loss, using precontrast geometry and an O-derivative induced by the loss. The framework includes maximum likelihood and posterior-based estimators and recovers existing bias-reduction methods based on the choice of prior or adjustments to estimating equations. In the likelihood setting, for the Jeffreys posterior μn and its Fisher--Rao Fréchet mean ẑn, the corrected estimator 2γ(ẑn)−μn(γ) has frequentist bias o(an2) for every fixed smooth functional γ, where an denotes the estimation rate. For coordinate squared loss, the theory yields a diffusion determined only by the likelihood and Fisher information. The mean of its invariant probability measure estimates the parameter with bias o(an2). This construction remains available even when no bias-reducing prior exists. Numerical results in the gamma shape--scale model illustrate this bias reduction.
Related
- Deep operator learning for efficient sampling from invariant measures of stochastic differential equations
- Invariant measures for the one-dimensional stochastic Navier-Stokes-Korteweg equations
- infomeasure-rs: A High-Performance Rust Library for Information-Theoretic Measures and Estimators
- Sharp bounds for stochastic proximal and projection estimators via radial dominance
- Mean squared error reduction of (plug-in) look-ahead estimators for Markov chains
BibTeX
@misc{yano2026invariant,
title = {{Invariant Measures as Estimators: Second-Order Stochastic Expansions and Bias Reduction}},
author = {Shota Yano},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.39355},
eprint = {2609.39355},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.39355},
}
RIS
TY - UNPB
TI - Invariant Measures as Estimators: Second-Order Stochastic Expansions and Bias Reduction
AU - Shota Yano
PY - 2026
DA - 2026/09/30
PB - arXiv
DO - 10.48550/arxiv.2609.39355
UR - https://arxiv.org/abs/2609.39355
AB - We develop a general second-order asymptotic framework for estimators constructed from invariant probability measures of data-dependent Markov processes. The framework yields a universal bias correction for smooth functionals, computed from the same invariant measure used to construct the estimator, without the resampling and repeated estimation involved in bias-correction methods such as the bootstrap and jackknife. The statistical input is represented by a random one-form, allowing differentials of log-likelihoods and general random criteria, as well as estimating functions, to be treated in a common framework. Under local regularity and localization conditions, we derive second-order stochastic expansions for averages with respect to these invariant measures and for decision rules defined through general loss functions. The expansions separate the contributions of the statistical input, temperature, drift, and loss, using precontrast geometry and an O-derivative induced by the loss. The framework includes maximum likelihood and posterior-based estimators and recovers existing bias-reduction methods based on the choice of prior or adjustments to estimating equations. In the likelihood setting, for the Jeffreys posterior μ_n and its Fisher--Rao Fréchet mean ẑ_n, the corrected estimator 2γ(ẑ_n)−μ_n(γ) has frequentist bias o(a_n^2) for every fixed smooth functional γ, where a_n denotes the estimation rate. For coordinate squared loss, the theory yields a diffusion determined only by the likelihood and Fisher information. The mean of its invariant probability measure estimates the parameter with bias o(a_n^2). This construction remains available even when no bias-reducing prior exists. Numerical results in the gamma shape--scale model illustrate this bias reduction.
ER -