.png){kind=icon}
返回
Dual-Path Modeling With Memory Embedding Model for Continuous Speech Separation
.png){kind=icon}
PRE
AI
DOI：10.1109/TASLP.2022.3165712.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Continuous speech separation (CSS) aims at separating overlap-free targets from a long, partially-overlapped recording. Though it has shown promising results, the origin CSS framework does not consider cross-window information and long-span dependency. To alleviate these limitations, this work introduces two novel methods to implicitly and explicitly capture the long-span knowledge, respectively. We firstly apply the dual-path (DP) modeling architecture for the CSS framework, where the within and across window information are jointly modeled by alternating stacked local-global processing modules. Secondly, to further capture the long-span dependency, we introduce a memory-based model for CSS. An additional memory pool is designed to extract embedding from each small window, and the inter-window commutation is established above the memory embedding pool through an attention mechanism. This memory-based model can precisely control what information needs to be transferred across the windows, thus leading to both improved modeling capacity and interpretability. The experimental results on the LibriCSS dataset show that both strategies can well capture the long-span information of the continuous speech and significantly improve system performance. Moreover, further improvements are observed with the integration of these two methods.
Keyword:
Recording
Speech processing
Computational modeling
Task analysis
Particle separators
Operating systems
Training
Continuous speech separation
dual-path modeling
memory pool
overlap ratio predictor
期刊
I
IF:
5.1
论文数:
2.6K
被引数:
1.1W
机构
引用论文
分享收藏Neonatal caffeine administration causes a permanent increase in the dendritic length of prefrontal cortical neurons of rats.png)
Synapse
SynapseIF0
PREAI
分享收藏Shape analysis of the neostriatum in frontotemporal lobar degeneration, Alzheimer's disease, and controlsNeuroImage
NeuroImageIF0
PREAI
分享收藏Cold event at 5 500 a BP recorded in mud sediments on the inner shelf of the East China Sea东中国海陆架内区泥质沉积物中记录的5500年前冷事件
PREAI
分享收藏
查看更多内容