arXiv Machine Learning

A Generalized Tangent Approximation based Variational Inference Framework for Strongly Super-Gaussian Likelihoods

The paper introduces a new variational inference framework that uses tangent transformations to handle strongly super‑Gaussian likelihoods across a wide range of probability models. By constructing tangent minorants of the log‑likelihood through convex duality, the method achieves conjugacy with Gaussian priors, enabling tractable inference where traditional approaches struggle. The authors provide algorithmic convergence guarantees and near‑parametric risk bounds, and demonstrate superior scalability and accuracy on both simulated and real‑world datasets compared to existing variational algorithms.

arXiv Machine Learning
Aug 28

A Flexible Empirical Bayes Approach to Generalized Linear Models, with Applications to Sparse Logistic Regression

The paper presents a tuning‑free empirical Bayes framework for Bayesian generalized linear models that uses a novel mean‑field variational inference algorithm. By estimating the prior within the VI procedure and optimizing the posterior mean directly, the method reduces optimization complexity and supports scalable solvers like L‑BFGS and stochastic gradient descent. Applied to sparse logistic regression, the approach shows superior predictive performance compared to existing methods.

By Dongyue Xie, Matthew Stephens
arXiv Statistics ML
6d ago

Scalable and Calibrated Sampling for Bayesian Generalized Linear Mixed Model via Stochastic Gradient Markov Chain Monte Carlo

The paper introduces a stochastic gradient Markov Chain Monte Carlo (SGMCMC) algorithm designed for Bayesian generalized linear mixed models (GLMMs). By employing Fisher's identity, it constructs a biased Monte Carlo estimator of the gradient of the marginal log-likelihood, enabling scalable inference even when direct gradient computation is infeasible. The authors provide a post-hoc covariance correction to calibrate posterior uncertainty and demonstrate the method’s accuracy and practical impact on large-scale biomedical data, notably electronic health records.

By Youngsoo Baek, Andrea Agazzi, Felipe A. Medeiros, Samuel I. Berchuck
arXiv Machine Learning
Aug 27

Gradient-based Sample Selection for Faster Bayesian Optimization

The paper introduces Gradient-based Sample Selection Bayesian Optimization (GSSBO), a method that builds the Gaussian process surrogate on a strategically chosen subset of samples rather than the full dataset. By using gradient information to eliminate redundant points while keeping diversity and representativeness, GSSBO achieves sublinear regret bounds and reduces the cubic computational cost of standard BO. Experiments on synthetic and real-world tasks show that this approach maintains comparable optimization performance while significantly cutting GP fitting time and resource usage.

By Qiyu Wei, Haowei Wang, Zirui Cao, Songhao Wang, Richard Allmendinger, Mauricio A \'Alvarez
arXiv Machine Learning
Sep 3

Neural Variational Cut Posteriors without Upstream Data

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
arXiv Machine Learning
Oct 16

Acceleration and its Limits in Gaussian Variational Inference

The paper investigates whether momentum-based acceleration can improve Gaussian variational inference (GVI) for strongly log-concave, log-smooth targets. With exact Gaussian expectations, GVI achieves the classic accelerated √κ convergence rate, but when only pointwise derivatives are available, the authors construct targets that force any randomized algorithm to use κ^{1-o(1)} queries, showing curvature alone does not guarantee acceleration. They identify additional structure—specifically a Hessian-Lipschitz condition with small Hessian variation—under which variance-reduced stochastic methods can achieve accelerated guarantees, and demonstrate empirical speedups on real datasets.

By Jaesung R. Park, Uijeong Jang, Yifan Chen