Maryclare Griffin

Miscellanea

Code

A mix of vignettes for little R packages I make as part of my research and statistical musings, i.e. things I thought about for a while but are not publishable or notes to myself that I thought could also be helpful to others.

Papers I Want To Read and Would Probably Love

1/4/2024-1/11/2024
- Jana, Fan, and Kulkarni, “A General Theory for Robust Clustering via Trimmed Mean” addresses the important problem of clustering when the data have Gaussian or heavier-than-Gaussian tails, which seems of great practical importance!
- Dobriban, Su, Yang, and Zhang, “Robust Inference Under Heteroskedasticity via the Hadamard Estimator” looks like a super interesting paper about properties of an alternative to Whit’s sandwich estimator!
- Sun, “Do we need to estimate the variance in robust mean estimation?” introduces a new method for obtaining an efficient estimator of the mean in the presence of heavy tails without having to estimate the unknown variance parameter, which is needed for tuning!
- Rodriguez-Alvarez, Inacio, and Klein, “Density regression via Dirichlet process mixtures of normal structured additive regression models” is an interesting paper introducing a new model that can relate the distribution of a continuous variable to covariates!
- Sanz-Alonso and Waniorek, “Hierarchical Bayesian Inverse Problems” is an interesting paper exploring the properties of the posterior mode for Bayesian penalized regression problems.
- Hallin and Konen, “Multivariate Quantiles: Geometric and Measure-Transportation-Based Contours” addresses the important problem of defining and computing quantiles of multivariate distributions!
- Craiu and Meng, “Perfecting MCMC Sampling: Recipes and Reservations” looks like a fantastic review of coupling methods in MCMC methodology!
- Pearse, Cressie, and Gunawan, “Optimal prediction of positive-valued spatial processes: asymmetric power-divergence loss” is an interesting paper about the use of alternative measures of loss when working with positive valued data, which considers log-Gaussian progesses as a special case!
12/31/2023-1/3/2024
- Deliu and Liseo, “Alternative Approaches for Computing Highest-Density Regions” addresses the important problem of computing highest-density regions for multivariate data, which seems relevant to characterizing the MCMC output in high dimensional problems!
- Korte-Stapff, Karvonen, and Moulines, “Smoothness Estimation for Whittle-Matern Processes on Closed Riemannian Manifolds” examines estimation of the smoothness parameter of the Matern covariance function, which is something I think is important!
- Soloff, Guntuboyina, and Sen, “Multivariate, Heteroscedastic Empirical Bayes via Nonparametric Maximum Likelihood” addresses the problem of multivariate mean estimation from heteroscedastic data using the nonparametric maximum likelihood estimator for Gaussian location mixtures, which is something I’ve been meaning to learn more about!
- Muehlebach and Jordan, “Accelerated First-Order Optimization under Nonlinear Constrants” introduces new accelerated first order algorithms that apply to $\ell_q$ penalized regression, which is of great interest to me!
- Kent, “Penalty Parameter Selection in Deconvolution by Estimating the Risk for a Smaller Sample Size” is an interesting paper about the deconvolution problem, which is a problem I would like to learn more about!
- Kubal, Campbell, and Robeva “Log-concave Density Estimation with Independent Components” addresses the challenging problem of performing log-concave density estimation for a multivariate density.
- Kanagawa, Barp, Gretton, and Mackey, “Controlling Moments with Kernel Stein Discrepencies” examines the performance of Kernel Stein discepancies, which are quantities that can be used to measure the quality of an approximation. Specifically, it addresses how/when Kernel Stein discrepancies can be interpreted in terms of convergence of moments of the distribution that is being approximated, which seems very valuable!