Research questionHow can we generate text-guided motion for arbitrary 3D skeletons without per-skeleton adaptation or reference motions?Motion generation must accommodate different kinematic structures while interpreting natural-language descriptions. Category-specific templates and skeleton-specific adaptation limit how broadly one animator can be applied. Latest papersRecent research connected to this question, newest first.UniMate: One Unified Model to Animate Diverse SkeletonsThe source concerns synthesizing articulated motion from a rigged 3D asset and a text prompt, with no test-time optimization or per-skeleton retraining. It reports training on 13,006 motion sequences covering animal forms and articulated rigid objects, along with zero-shot cross-topology transfer, in-betweening, motion expansion, and text-guided editing; the supplied evidence reports improvements in quality, generalization, and efficiency but does not detail the evaluation protocols.research paper · Sep 4, 2026