Yanmin Qian

TF-MoE achieves a remarkable +3.8 dB improvement in speech separation performance while keeping computational costs low, making it ideal for edge-device applications.

LLMs can now evaluate audio as well as humans, without task-specific training, thanks to a new instruction-driven framework.

Aligning speech VAEs with SSL features isn't a one-size-fits-all game: joint-marginal alignment with adaptive weighting unlocks superior performance across reconstruction, understanding, and generation.