PhD student, The Chinese University of Hong Kong, Shenzhen
1 paper at NeurIPS 2025
A novel speech tokenizer with an end-to-end diffusion autoencoder and text-aware decoding, operating at 6.25 Hz and 0.0875 kbps
