Probing Self-Supervised Learning Models With Target Speech Extraction

conference paper

English

Large-scale pre-trained self-supervised learning (SSL) models have shown remarkable advancements in speech-related tasks. However, the utilization of these models in complex multi-talker scenarios, such as extracting a target speaker in a mixture, is yet to be fully evaluated. In this paper, we introduce target speech extraction (TSE) as a novel downstream task to evaluate the feature extraction capabilities of pre-trained SSL models. TSE uniquely requires both speaker identification and speech separation, distinguishing it from other tasks in the Speech processing Universal PERformance Benchmark (SUPERB) evaluation. Specifically, we propose a TSE downstream model composed of two lightweight task-oriented modules based on the same frozen SSL model. One module functions as a speaker encoder to obtain target speaker information from an enrollment speech, while the other estimates the target speaker's mask to extract its speech from the mixture. Experimental results on the Libri2mix datasets reveal the relevance of the TSE downstream task to probe SSL models, as its performance cannot be simply deduced from other related tasks such as speaker verification and separation.

Target speech extraction, self-supervised learning, SUPERB

PENG, J.; DELCROIX, M.; OCHIAI, T.; ASHIHARA, T.; PLCHOT, O.; ARAKI, S.; ČERNOCKÝ, J.

14. 4. 2024

IEEE Signal Processing Society

Seoul

979-8-3503-7451-3

ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings

535

539

5

https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10627502

peng_icassp2024_Probing_Self-Supervised_Learning.pdf