1 paper
Hao Yen, Shaoshi Ling, Guoli Ye
We propose a novel approach to end-to-end automatic speech recognition (ASR) to achieve efficient speech in-context learning (SICL) for (i) long-form speech decoding, (ii) test-tim…