The STC System for the CHiME-6 Challenge
The STC System for the CHiME-6 Challenge
复制标题
用于 CHiME-6 挑战的 STC 系统
DOI：
--
复制
发表时间：
2020
期刊：
影响因子：
--
通讯作者：
A. Romanenko
中科院分区：
文献类型：
--
作者：
I. Medennikov;M. Korenevsky;Tatiana Prisyach;Yuri Y. Khokhlov;Mariya Korenevskaya;Ivan Sorokin;Tatiana Timofeeva;Anton Mitrofanov;A. Andrusenko;Ivan Podluzhny;A. Laptev;A. Romanenko
本文描述了针对CHINE-6挑战的语音技术中心(STC)系统，该挑战旨在晚宴场景中的多麦克风、多说话人语音识别和二元化。我们参加了赛道1和赛道2，并提交了每个赛道的排名A和排名B的结果。一轨系统采用了基于软活动的导引源分离(GSS)作为前端，结合了基于GSS的训练数据增强、多步长多流自关注层、统计层和谱增强等先进的声学建模技术，以及声学模型的格子级融合。我们的Track 1系统排在前三位，在基线的基础上实现了30%的相对WER降低。此外，使用神经语言模型对B进行格点重新排序。总体来说，这导致轨道1的相对WER比基线降低了34%。对于轨道2，我们提出了一种新的目标-说话人语音活动检测(TS-VAD)方法来解决二值化问题。良好的二值化结果使得在获得的片段上执行GSS成为可能。TS-VAD基于i向量说话人嵌入，最初使用基于x向量谱聚类的强二值化系统进行估计。第二个轨道使用来自轨道1系统的后端。赛道2的系统表现出了最先进的性能，相对地超过了基线39%的DER、45%的JER、43%的WER(排名A)和45%的WER(排名B)。
This paper is a description of the Speech Technology Center (STC) systems for the CHiME-6 challenge aimed at multi-microphone multi-speaker speech recognition and diarization in a dinner party scenario. We participated in both Track 1 and Track 2 and submitted our results for Ranking A as well as Ranking B for each track. The soft-activity based Guided Source Separation (GSS) as a front-end and a combination of advanced acoustic modeling techniques such as GSS-based training data augmentation, multi-stride and multi-stream self-attention layers, statistics layer and SpecAugment, as well as the lattice-level fusion of acoustic models were applied in the 1st track system. Our system for Track 1 was in the top three systems, achieving 30% relative WER reduction over the baseline. Additionally, lattice rescoring with a neural language model was applied for Ranking B. Overall, this led to 34% relative WER reduction over the baseline in Track 1. For Track 2, we proposed a novel Target-Speaker Voice Activity Detection (TS-VAD) approach to solve the diarization problem. Good diarization results made it possible to perform GSS on the obtained segments. TS-VAD is based on i-vector speaker embeddings, which are initially estimated using a strong diarization system based on spectral clustering of x-vectors. The back-end from the Track 1 system was used in the second track. The system for Track 2 demonstrated state-of-the-art performance, outperforming the baseline by 39% DER, 45% JER, 43% WER (Ranking A) and 45% WER (Ranking B) relative.