Direct Preference Optimization for English-Mandarin Code-Switching Speech Recognition in Audio LLMs
TL;DR
Direct Preference Optimization (DPO) is applied to align models, constructing preference pairs in which chosen responses preserve mixed-language content while rejected responses mimic failure patterns, suggesting DPO can effectively elicit correct code-switching transcription behavior from multilingual Audio LLMs.
Abstract
Audio large language models (Audio LLMs) exhibit systematic failures in transcribing code-switching speech despite strong multilingual capabilities. Focusing on English-Mandarin, we identify three failure modes: language omission, translation-instead-of-transcription, and hallucination. We apply Direct Preference Optimization (DPO) to align models, constructing preference pairs in which chosen responses preserve mixed-language content while rejected responses mimic failure patterns. Training three Audio LLMs on 100K pairs (570 hours), we observe consistent behavioral shifts: models learn to preserve language composition rather than translating when prompted for transcription. This alignment yields MER reductions up to 89.6% (in-distribution) and 20.0% (out-of-distribution). Our findings suggest DPO can effectively elicit correct code-switching transcription behavior from multilingual Audio LLMs.
