Song, Z., Zhang, R., Li, S., Liu, C. (2026). Contrastive learning-driven spatiotemporal dynamically adaptive framework for stylized 3D human motion generation. PLoS ONE. https://doi.org/10.1371/journal.pone.0337596