Neural Architecture Search for Speech Emotion Recognition

TL;DR

This paper applies neural architecture search (NAS) to speech emotion recognition, automatically designing neural networks that improve performance while reducing manual effort and resource consumption.

Contribution

It introduces a novel NAS application for SER and proposes a uniform path dropout strategy to enhance architecture optimization.

Findings

NAS improves SER accuracy from 54.89% to 56.28%.

The dropout strategy outperforms previous methods.

Models maintain parameter sizes while improving performance.

Abstract

Deep neural networks have brought significant advancements to speech emotion recognition (SER). However, the architecture design in SER is mainly based on expert knowledge and empirical (trial-and-error) evaluations, which is time-consuming and resource intensive. In this paper, we propose to apply neural architecture search (NAS) techniques to automatically configure the SER models. To accelerate the candidate architecture optimization, we propose a uniform path dropout strategy to encourage all candidate architecture operations to be equally optimized. Experimental results of two different neural structures on IEMOCAP show that NAS can improve SER performance (54.89\% to 56.28\%) while maintaining model parameter sizes. The proposed dropout strategy also shows superiority over the previous approaches.