arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 2412.08548cs.CL
收藏

用于自动语音识别的双层无监督与监督联合训练

Bilevel Joint Unsupervised and Supervised Training for Automatic Speech Recognition

IBM Research(IBM研究院);IBM T. J. Watson Research Center(IBM T. J. 沃森研究中心);Rensselaer Polytechnic Institute(伦斯勒理工学院)

Xiaodong Cui, A F M Saif, Songtao Lu, Lisha Chen, Tianyi Chen, Brian Kingsbury, George Saon

AI总结:

该研究提出用于自动语音识别的BL-JUST双层联合训练框架,通过基于惩罚的双层梯度下降同时优化无监督与监督损失,使声学表示兼顾通用性与任务特异性,性能优于传统预训练微调及主流半监督方法。

摘要:

本文提出一种用于自动语音识别的双层无监督与监督联合训练(BL-JUST)框架。与传统的预训练加微调这种相互脱节的两阶段策略不同,BL-JUST 旨在优化声学模型,使其同时最小化无监督损失函数与监督损失函数。由于 BL-JUST 寻求两类损失函数的匹配局部最优解,声学模型学习到的声学表示能在通用性与任务特异性之间取得良好平衡。我们采用基于惩罚的双层梯度下降法求解 BL-JUST 问题,并在多种数据集上,结合不同架构与损失函数评估训练得到的深度神经网络声学模型。实验表明,BL-JUST 的性能优于广泛使用的预训练微调策略及其他一些主流半监督技术。

英文摘要:

In this paper, we propose a bilevel joint unsupervised and supervised training (BL-JUST) framework for automatic speech recognition. Compared to the conventional pre-training and fine-tuning strategy which is a disconnected two-stage process, BL-JUST tries to optimize an acoustic model such that it simultaneously minimizes both the unsupervised and supervised loss functions. Because BL-JUST seeks matched local optima of both loss functions, acoustic representations learned by the acoustic model strike a good balance between being generic and task-specific. We solve the BL-JUST problem using penalty-based bilevel gradient descent and evaluate the trained deep neural network acoustic models on various datasets with a variety of architectures and loss functions. We show that BL-JUST can outperform the widely-used pre-training and fine-tuning strategy and some other popular semi-supervised techniques.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑