Evan Gorstein

This work introduces the first framework that generates plausible ancestral vocalizations, and shows that this method is the only approach that simultaneously achieves genuine generation, phylogenetic consistency, and naturalistic audio quality across both datasets.