English

Multimodal Large Language Models with Fusion Low Rank Adaptation for Device Directed Speech Detection

Computation and Language 2024-06-17 v1 Human-Computer Interaction Audio and Speech Processing

Abstract

Although Large Language Models (LLMs) have shown promise for human-like conversations, they are primarily pre-trained on text data. Incorporating audio or video improves performance, but collecting large-scale multimodal data and pre-training multimodal LLMs is challenging. To this end, we propose a Fusion Low Rank Adaptation (FLoRA) technique that efficiently adapts a pre-trained unimodal LLM to consume new, previously unseen modalities via low rank adaptation. For device-directed speech detection, using FLoRA, the multimodal LLM achieves 22% relative reduction in equal error rate (EER) over the text-only approach and attains performance parity with its full fine-tuning (FFT) counterpart while needing to tune only a fraction of its parameters. Furthermore, with the newly introduced adapter dropout, FLoRA is robust to missing data, improving over FFT by 20% lower EER and 56% lower false accept rate. The proposed approach scales well for model sizes from 16M to 3B parameters.

Keywords

Cite

@article{arxiv.2406.09617,
  title  = {Multimodal Large Language Models with Fusion Low Rank Adaptation for Device Directed Speech Detection},
  author = {Shruti Palaskar and Oggi Rudovic and Sameer Dharur and Florian Pesce and Gautam Krishna and Aswin Sivaraman and Jack Berkowitz and Ahmed Hussen Abdelaziz and Saurabh Adya and Ahmed Tewfik},
  journal= {arXiv preprint arXiv:2406.09617},
  year   = {2024}
}

Comments

Accepted at Interspeech 2024