A1455
Title: Non-asymptotic Laplace approximation under model misspecification
Authors: Anirban Bhattacharya - Texas AM University (United States) [presenting]
Debdeep Pati - Texas A&M University (United States)
Abstract: Non-asymptotic two-sided bounds are presented to the log-marginal likelihood in Bayesian inference. The classical Laplace approximation is recovered as the leading term. The derivation permits model misspecification and allows the parameter dimension to grow with the sample size. No assumptions are made about the asymptotic shape of the posterior, and instead, certain regularity conditions on the likelihood ratio and that the posterior is sufficiently concentrated. The derived bounds are envisioned to be widely applicable in establishing model selection consistency of Bayesian procedures in non-conjugate settings, especially when the true model potentially lies outside the class of candidate models considered.
