Wang-Jin Lim

STEM-LM uses masked language modeling to predict multi-species distributions by treating species presence as tokens in a spatio-temporal 'sentence', achieving high discrimination on ecological datasets.

The paper corrects an overcounting issue in the unresolved FBD model for tip-dating and demonstrates its application to empirical datasets.

The paper introduces a hierarchical Bayesian framework to estimate taxon-specific intraspecific covariance structures and evolutionary processes for multivariate traits, showing improved uniformity in evolutionary rate estimates across the great ape clade.

DICAROS improves ancestral shape reconstruction by modeling shape correlations using diffeomorphic dynamics, outperforming traditional methods on non-symmetric trees.