Search papers, labs, and topics across Lattice.
Huawei Technology,China, University of Science and Technology of China,China, University of Science and Technology of China
2
0
4
0
Synthesizing realistic dysarthric speech slashes ASR error rates by over 50%, thanks to a novel rhythm-style modeling approach.
Achieve near-real-time dysarthric speech reconstruction with a 54% WER reduction by jointly training ASR and TTS with a novel frame-level adaptor and knowledge distillation.
