How to Estimate Model Transferability of Pre-Trained Speech Models?

Zih-Ching Chen, Chao-Han Huck Yang, Bo Li, Yu Zhang, Nanxin Chen, Shuo-Yiin Chang, Rohit Prabhavalkar,Hung-yi Lee, Tara N. Sainath

CoRR(2023)

引用 0|浏览67
暂无评分
摘要
In this work, we introduce a "score-based assessment" framework for estimating the transferability of pre-trained speech models (PSMs) for fine-tuning target tasks. We leverage upon two representation theories, Bayesian likelihood estimation and optimal transport, to generate rank scores for the PSM candidates using the extracted representations. Our framework efficiently computes transferability scores without actual fine-tuning of candidate models or layers by making a temporal independent hypothesis. We evaluate some popular supervised speech models (e.g., Conformer RNN-Transducer) and self-supervised speech models (e.g., HuBERT) in cross-layer and cross-model settings using public data. Experimental results show a high Spearman's rank correlation and low p-value between our estimation framework and fine-tuning ground truth. Our proposed transferability framework requires less computational time and resources, making it a resource-saving and time-efficient approach for tuning speech foundation models.
更多
查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络