Bomin Liu, Linjun He, Yan Zhu. "MViT: A vision transformer with fractal path reordering and dynamic positional encoding." PLoS ONE (2026). https://doi.org/10.1371/journal.pone.0340788