Multimodal Large Language Models with Fusion Low Rank Adaptation for Device Directed Speech Detection
This work addresses the problem of efficient multimodal adaptation for device-directed speech detection, offering a parameter-efficient solution with robustness to missing data, though it is incremental in improving existing adaptation methods.
The paper tackles the challenge of adapting pre-trained unimodal large language models to multimodal tasks like device-directed speech detection, proposing a Fusion Low Rank Adaptation (FLoRA) technique that achieves a 22% relative reduction in equal error rate over text-only methods and matches full fine-tuning performance with fewer parameters.
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.