TY  - RPRT
TI  - MindVoice: Reconstructing Intelligible Speech from Non-invasive Neural Signals with Pretrained Priors
AU  - Guangyin Bao
AU  - Taiping Zeng
AU  - Jianfeng Feng
AU  - Xiangyang Xue
PY  - 2026
UR  - https://arxiv.org/abs/2605.31173
ID  - 2605.31173
ER  -