Search papers, labs, and topics across Lattice.
1
0
3
Reinforcement learning with audio-text semantic rewards can overcome confirmation bias in test-time adaptation, leading to more robust ASR in noisy and accented speech environments.
