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 2408.03440eess.AScs.SD
收藏

TF-Locoformer:用于语音分离与增强的基于卷积局部建模的Transformer

TF-Locoformer: Transformer with Local Modeling by Convolution for Speech Separation and Enhancement

MERL(三菱电机美国研究实验室)

Kohei Saijo, Gordon Wichern, François G. Germain, Zexu Pan, Jonathan Le Roux

AI总结:

本文提出TF-Locoformer,用卷积前馈网络替代RNN实现时频域双路径语音分离与增强,在保持局部建模能力的同时达到SoTA性能。

摘要:

时频(TF)域双路径模型实现了高保真语音分离。虽然一些先前的最先进(SoTA)模型依赖RNN,但这种依赖意味着它们缺乏Transformer模块的并行性、可扩展性和通用性。鉴于纯Transformer架构在其他领域的广泛成功,本文专注于从TF域双路径模型中移除RNN,同时保持SoTA性能。本文提出了TF-Locoformer,一种基于Transformer并通过卷积进行局部建模的模型。该模型使用带有卷积层的前馈网络(FFN)代替线性层来捕获局部信息,使自注意力专注于捕获全局模式。我们在自注意力前后各放置两个这样的FFN,以增强局部建模能力。我们还为TF域双路径模型引入了一种新颖的归一化方法。在分离和增强数据集上的实验表明,所提出的模型在无RNN架构下,在多个基准上达到或超过了SoTA性能。

英文摘要:

Time-frequency (TF) domain dual-path models achieve high-fidelity speech separation. While some previous state-of-the-art (SoTA) models rely on RNNs, this reliance means they lack the parallelizability, scalability, and versatility of Transformer blocks. Given the wide-ranging success of pure Transformer-based architectures in other fields, in this work we focus on removing the RNN from TF-domain dual-path models, while maintaining SoTA performance. This work presents TF-Locoformer, a Transformer-based model with LOcal-modeling by COnvolution. The model uses feed-forward networks (FFNs) with convolution layers, instead of linear layers, to capture local information, letting the self-attention focus on capturing global patterns. We place two such FFNs before and after self-attention to enhance the local-modeling capability. We also introduce a novel normalization for TF-domain dual-path models. Experiments on separation and enhancement datasets show that the proposed model meets or exceeds SoTA in multiple benchmarks with an RNN-free architecture.

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。

↑