Zhaoyang Li

HiCoDiT achieves superior audio-visual alignment by harnessing the hierarchical nature of speech tokens, outperforming traditional VTS methods in both fidelity and expressiveness.

Soft pseudo-labels, theoretically equivalent to hard labels when perfectly calibrated, tank performance in cross-domain semantic segmentation, motivating a new calibration framework.

LLMs can translate natural-language user intents into executable physical-layer control decisions in 6G wireless, enabling more autonomous and adaptable communication systems.