Resurfacing Paralinguistic Awareness in Large Audio Language Models
This work addresses the limitation of LALMs in ignoring paralinguistic cues for human interaction, which is an incremental improvement for audio-based AI systems.
The paper tackles the problem of Large Audio Language Models (LALMs) neglecting paralinguistic cues by proposing a paralinguistic-enhanced fine-tuning (PE-FT) protocol, which includes selective-layer fine-tuning and an auxiliary dual-level classification head, and demonstrates that this protocol efficiently resurfaces paralinguistic awareness, surpassing all-layer fine-tuning in performance.
Large Audio Language Models (LALMs) have expanded the interaction with human to speech modality, which introduces great interactive potential, due to the paralinguistic cues implicitly indicating the user context. However, building on the current content-centred paradigm, LALMs usually neglect such paralinguistic cues and respond solely based on query content. In this work, to resurface the paralinguistic awareness in LALMs, we introduce five diverse layer-wise analyses to jointly identify paralinguistic layers and semantic understanding layers. Based on these insights, we propose a paralinguistic-enhanced fine-tuning (PE-FT) protocol accordingly to equip LALMs with paralinguistic-aware capabilities, including (1) selective-layer fine-tuning, and (2) an auxiliary dual-level classification head. Our experiments demonstrate that PE-FT protocol efficiently and effectively resurfaces the paralinguistic awareness, even surpassing the performance of the all-layer fine-tuning strategy.