Multimodal Attention Merging for Improved Speech Recognition and Audio Event Classification
Training large foundation models using self-supervised objectives on unlabeled data, followed by fine-tuning on downstream tasks, has emerged as a standard procedure. Unfortunately, the efficacy of this approach is often constrained by both limited fine-tuning compute and scarcity in labeled downstr...
Saved in:
Published in | 2024 IEEE International Conference on Acoustics, Speech, and Signal Processing Workshops (ICASSPW) pp. 655 - 659 |
---|---|
Main Authors | , , , , , |
Format | Conference Proceeding |
Language | English |
Published |
IEEE
14.04.2024
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | Training large foundation models using self-supervised objectives on unlabeled data, followed by fine-tuning on downstream tasks, has emerged as a standard procedure. Unfortunately, the efficacy of this approach is often constrained by both limited fine-tuning compute and scarcity in labeled downstream data. We introduce Multimodal Attention Merging (MAM), an attempt that facilitates direct knowledge transfer from attention matrices of models rooted in high-resource modalities, text and images, to those in resource-constrained domains, speech and audio, employing a zero-shot paradigm. MAM reduces the relative Word Error Rate (WER) of an Automatic Speech Recognition (ASR) model by up to 6.70%, and relative classification error of an Audio Event Classification (AEC) model by 10.63%. In cases where some data/compute is available, we present Learnable-MAM, a data-driven approach to merging attention matrices, resulting in a further 2.90% relative reduction in WER for ASR and 18.42% relative reduction in AEC compared to fine-tuning without model merging. |
---|---|
DOI: | 10.1109/ICASSPW62465.2024.10627466 |