A Layer-Anchoring Strategy for Enhancing Cross-Lingual Speech Emotion Recognition

TL;DR

This paper introduces a layer-anchoring strategy that leverages hierarchical features across transformer layers to improve cross-lingual speech emotion recognition, demonstrating significant performance gains on multiple datasets.

Contribution

The study proposes a novel layer-anchoring mechanism that utilizes information from different transformer layers to enhance emotion transfer in cross-lingual SER tasks.

Findings

Achieved up to 60.21% UAR on BIIC-podcast corpus.

Layer feature similarity varies across languages and layers.

Layer-anchoring improves cross-lingual emotion recognition performance.

Abstract

Cross-lingual speech emotion recognition (SER) is important for a wide range of everyday applications. While recent SER research relies heavily on large pretrained models for emotion training, existing studies often concentrate solely on the final transformer layer of these models. However, given the task-specific nature and hierarchical architecture of these models, each transformer layer encapsulates different levels of information. Leveraging this hierarchical structure, our study focuses on the information embedded across different layers. Through an examination of layer feature similarity across different languages, we propose a novel strategy called a layer-anchoring mechanism to facilitate emotion transfer in cross-lingual SER tasks. Our approach is evaluated using two distinct language affective corpora (MSP-Podcast and BIIC-Podcast), achieving a best UAR performance of 60.21% on the BIIC-podcast corpus. The analysis uncovers interesting insights into the behavior of popular pretrained models.