English-Mandarin

DPO improves code-switching speech recognition in audio LLMs

Researchers have developed a new method using Direct Preference Optimization (DPO) to improve how audio large language models handle speech that switches between English and Mandarin. The models often fail by omitting l…