statistical inference

Efficient and Accurate Inference for Matérn Gaussian Processes on Intervals

1 min read ·
A new study by researchers at KAUST introduces a breakthrough method for statistical modeling with Matérn Gaussian processes, a popular tool in spatial statistics, machine learning, and the natural sciences. While Gaussian processes with stationary Matérn covariance functions (Matérn processes) are valued for their flexibility, their use with large datasets has been severely limited by the high computational cost of standard methods. The team has developed the first generally applicable approach that enables fast, linear-cost inference and prediction for Matérn processes on bounded intervals

High-dimensional Q-learning for Dynamic Treatment Regimes

-

B3 L5 R5220

Dynamic treatment regimes are a set of decision rules and each treatment decision is tailored over time according to patients’ responses to previous treatments as well as covariate history. There is a growing interest in development of correct statistical inference for optimal dynamic treatment regimes to handle the challenges of nonregularity problems in the presence of nonrespondents who have zero-treatment effects, especially when the dimension of the tailoring variables is high.

Transparent and Robust Causal Inference in Data Science

Carlos Cinelli, Ph.D. candidate, Department of Statistics, UCLA

-

KAUST

The past few decades have witnessed rapid and unprecedented theoretical progress in the science of causal inference, ranging from the “credibility revolution” with the popularization of quasi-experimental designs, to the development of a complete solution to non-parametric identification with causal graphical models. Most of these theoretical progress, however, relies on strong, exact assumptions, such as the absence of unobserved common causes, or the absence of certain direct effects. Unfortunately, more often than not these assumptions are very hard to defend in practice. This leads to two undesirable consequences for applied quantitative work in the data-intensive sciences: (i) important research questions may be neglected, simply because they do not exactly match the requirements of current methods; or, (ii) researchers may succumb to making the required “identification assumptions” simply to justify the use of available methods, but not because these assumptions are truly believed (or understood). In this talk, I will discuss new theories, methods, and software for permitting causal inferences under more flexible and realistic settings. These tools empower scientists, and policymakers to both examine the sensitivity of causal inferences to violations of its underlying assumptions, and also to draw robust and trustworthy conclusions from settings in which traditional methods fail.

Kelvin J. R. Almeida-Sousa

Postdoctoral Research Fellow, Statistics
Kelvin J. R. Almeida-Sousa is a Brazilian mathematician with contributions to deterministic and stochastic partial differential equations, numerical analysis, mathematical statistics, and spectral theory.

Alexandre de Bustamante Simas

Senior Research Scientist, Statistics
Alexandre Simas is an expert on stochastic processes and spatial statistics, with a focus on stochastic partial differential equations and random fields on metric graphs.