News | Somjit Roy

2026

Feb Work on Frequentist Regret Analysis of Gaussian Process Thompson Sampling via Fractional Posteriors submitted.
Jan Invited talk on Hierarchical Bayesian Operator-induced Symbolic Regression Trees for Structural Learning of Scientific Expressions (arXiv:2509.19710) at the 2026 Symposium on Data Science & Statistics in Milwaukee, Wisconsin.
Jan Our work Hierarchical Bayesian Operator-induced Symbolic Regression Trees for Structural Learning of Scientific Expressions (arXiv:2509.19710) was awarded the 2026 Section on Bayesian Statistical Science (SBSS) Student Paper Award by the American Statistical Association.

2025

Nov Preprint on Tangent approximation based variational inference for strongly super-Gaussian likelihoods submitted (arXiv).
Nov GitHub release of TAVIE-SSG.
Oct Talk on Bayesian Symbolic Regression for Structural Learning of Scientific Expressions at STAT CAFÉ, Texas A&M University.
Sep Invited talk on SPINWAVE at the minisymposium Computational Methods for Multiscale and Multiphysics Problems during the 8th Annual SIAM Texas–Louisiana Workshop, Austin, TX.
Sep Preprint on Bayesian Scientific Machine Learning for Structural Learning of Scientific Expressions submitted (arXiv).
Sep First official GitHub release of HierBOSSS.
May Graduate Intern in the Earth and Environmental Sciences (EES) Division at Los Alamos National Laboratory.

2024

Dec Talk on TAVIE: Tangent Approximation for Variational Inference in Different Exponential Families at IISA 2024.
Nov Paper on Sparse Almost Perfect Mutually Unbiased Bases (APMUBs) published in the Journal of Statistical Theory and Practice (Journal).