Spatial Attention for Far-field Speech Recognition with Deep Beamforming Neural Networks

TL;DR

This paper introduces a spatial attention mechanism for neural beamformers in far-field speech recognition, effectively focusing on relevant features from multiple directions to improve recognition accuracy.

Contribution

The paper proposes a novel spatial attention subnet that enhances neural beamformers by weighting features from different directions, reducing redundancy and improving speech recognition performance.

Findings

Up to 9% relative word error rate reduction

Spatial attention improves feature relevance

Enhanced neural beamformer performance

Abstract

In this paper, we introduce spatial attention for refining the information in multi-direction neural beamformer for far-field automatic speech recognition. Previous approaches of neural beamformers with multiple look directions, such as the factored complex linear projection, have shown promising results. However, the features extracted by such methods contain redundant information, as only the direction of the target speech is relevant. We propose using a spatial attention subnet to weigh the features from different directions, so that the subsequent acoustic model could focus on the most relevant features for the speech recognition. Our experimental results show that spatial attention achieves up to 9% relative word error rate improvement over methods without the attention.