Search papers, labs, and topics across Lattice.
1
0
3
Reinforcement learning outperforms supervised fine-tuning in adapting ASR systems to synthetic speech, achieving a 40% reduction in word error rates.
