arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2407.11641cs.SDeess.AS
收藏

探究标签拓扑与训练准则对ASR性能和对齐质量的影响

Investigating the Effect of Label Topology and Training Criterion on ASR Performance and Alignment Quality

RWTH Aachen University(亚琛工业大学);AppTek GmbH(AppTek 公司)

Tina Raissi, Christoph Lüscher, Simon Berger, Ralf Schlüter, Hermann Ney

AI总结:

本研究对比了不同标签拓扑和训练准则的ASR模型,评估对齐质量与性能,实验于LibriSpeech和Switchboard数据集进行。

摘要:

当前自动语音识别(ASR)的研究格局设想端到端方法与经典模块化系统之间存在明确划分。尽管这两种方法在需求及(不)优势方面的高层次比较通常已有讨论,但在相似条件下进行更细致的比较在文献中并不常见。在本工作中,我们提出了一项聚焦于标签拓扑和训练准则的比较。我们比较了两种判别性对齐模型,分别采用隐马尔可夫模型(HMM)和连接主义时间分类拓扑,以及两种利用因子化HMM和严格单调循环神经网络转换器的二阶标签上下文ASR模型。我们使用不同的度量来评估对齐质量,并比较我们最佳系统的词错误率和实时因子。实验在LibriSpeech 960小时和Switchboard 300小时任务上进行。

英文摘要:

The ongoing research scenario for automatic speech recognition (ASR) envisions a clear division between end-to-end approaches and classic modular systems. Even though a high-level comparison between the two approaches in terms of their requirements and (dis)advantages is commonly addressed, a closer comparison under similar conditions is not readily available in the literature. In this work, we present a comparison focused on the label topology and training criterion. We compare two discriminative alignment models with hidden Markov model (HMM) and connectionist temporal classification topology, and two first-order label context ASR models utilizing factored HMM and strictly monotonic recurrent neural network transducer, respectively. We use different measurements for the evaluation of the alignment quality, and compare word error rate and real time factor of our best systems. Experiments are conducted on the LibriSpeech 960h and Switchboard 300h tasks.

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。

↑