The paper investigates how variational inference using normalizing flows can struggle with multimodal posteriors, particularly when modes are separated by low-probability regions. By adding a Rényi divergence-based regularization term, the authors encourage the approximate posterior to cover the prior support, reducing mode collapse from 50% to 25% in a twin experiment. However, this regularization alters the Bayesian objective and introduces a trade‑off in approximation accuracy, and it does not solve mode collapse in general.
arXiv:2512. 04954v3 Announce Type: replace Abstract: We present a novel technique for amortized posterior estimation using Normalizing Flows trained with likelihood-weighted importance sampling.
By Rajneil Baruah
arXiv:2606. 01954v1 Announce Type: new Abstract: Implicit-process priors define distributions over functions through flexible generative mechanisms, making them attractive for Bayesian function-space modelling.
By Luis A. Ortega, Andr\'es R. Masegosa, Thomas D. Nielsen
The paper introduces a method for localized posterior inference that weights observation likelihoods based on proximity to a target context, using Gaussian localization and a conditional normalizing flow for amortized variational inference. By integrating these localized posteriors over a context measure, the authors obtain a marginalized parameter distribution that accounts for observation and context uncertainties. Experiments with the Branin-Hoo model and a k‑ε turbulence closure in an ocean model demonstrate that localization reduces aggregate predictive errors and improves temperature and zonal‑velocity predictions compared to global or deterministic calibration.
By Exauc\'e Luweh Adjim Ngarti (AIRSEA), \'Elise Arnaud (AIRSEA), L\'eo Nicoletti (Mines Saint-\'Etienne MSE, LIMOS, FAYOL-ENSMSE, FAYOL-ENSMSE), Victor Trappler (Mines Saint-\'Etienne MSE, LIMOS, FAYOL-ENSMSE, FAYOL-ENSMSE), Arthur Vidard (AIRSEA)
arXiv:2606. 25882v1 Announce Type: new Abstract: DGPs are probabilistic models with remarkable prediction performance that concatenate GPs across several layers.
By Francisco Javier S\'aez-Maldonado, Juan Maro\~nas, Daniel Hern\'andez-Lobato
The paper introduces NeVI‑Cut, a modular variational inference method for cut‑Bayes that does not require access to upstream data or models. It approximates the cut‑posterior by minimizing the expected downstream conditional Kullback‑Leibler divergence, using conditional normalizing flows as the variational family. The authors provide fixed‑data convergence rates, establish uniform KL approximation results for flow classes, and demonstrate the algorithm’s speed and accuracy on several applications.
By Jiafang Song, Sandipan Pramanik, Abhirup Datta