2606.11033v1 Jun 09, 2026 cs.LG

AuRA: Internalizing Audio Understanding into LLMs as LoRA

Jun Xu
Jun Xu
Citations: 12
h-index: 2
Jiuchong Gao
Jiuchong Gao
Citations: 14
h-index: 2
Jinghua Hao
Jinghua Hao
Citations: 50
h-index: 2
Renqing He
Renqing He
Citations: 338
h-index: 9
Zhanyu Ma
Zhanyu Ma
Citations: 12
h-index: 2
Yuan Wu
Yuan Wu
Citations: 158
h-index: 5
Bo Cheng
Bo Cheng
Citations: 23
h-index: 3
Lei Shi
Lei Shi
Citations: 19
h-index: 3

Recent efforts to extend large language models (LLMs) to speech inputs typically rely on cascaded ASR-LLM pipelines, end-to-end speech-language models, or bridge/distillation-based adaptation. While these routes respectively reuse strong pretrained components, enable native speech-language interaction, or offer lightweight adaptation, they often suffer from transcript-interface latency, costly multimodal training, or sequential speech-language coupling. To address these limitations, we present AuRA, a method that distills audio encoding capability into the LLM. Specifically, AuRA feeds the same speech input to an ASR encoder (as a teacher) and a LoRA-adapted LLM (as a student) through a lightweight audio embedding layer, and uses layer-wise distillation to align the student's hidden states with corresponding teacher representations, thereby internalizing speech representations into lightweight LLM-side adaptations. Compared with cascaded and serial bridge methods, AuRA enables tighter speech-language joint modeling and efficient parallel end-to-end inference, while also reusing pretrained speech and language models rather than requiring large-scale multimodal training. On multiple speech-language benchmarks, AuRA consistently outperforms cascaded systems, speech-to-LLM adaptation baselines, and large-scale speech-language and multimodal models in both effectiveness and efficiency.

0 Citations
0 Influential
4.5 Altmetric
22.5 Score
Original PDF

No Analysis Report Yet

This paper hasn't been analyzed by Gemini yet.

Log in to request an AI analysis.

댓글

댓글을 작성하려면 로그인하세요.

아직 댓글이 없습니다. 첫 번째 댓글을 남겨보세요!