Contrastive Unsupervised Learning for Speech Emotion Recognition
This work addresses the lack of labeled datasets for speech emotion recognition, enabling more natural human-machine communication, though it is incremental as it applies an existing method to a specific domain.
The paper tackled the problem of limited labeled data in speech emotion recognition by using contrastive predictive coding for unsupervised representation learning, achieving state-of-the-art performance on IEMOCAP with improved concordance correlation coefficients and significant gains on the MSP-Podcast dataset.
Speech emotion recognition (SER) is a key technology to enable more natural human-machine communication. However, SER has long suffered from a lack of public large-scale labeled datasets. To circumvent this problem, we investigate how unsupervised representation learning on unlabeled datasets can benefit SER. We show that the contrastive predictive coding (CPC) method can learn salient representations from unlabeled datasets, which improves emotion recognition performance. In our experiments, this method achieved state-of-the-art concordance correlation coefficient (CCC) performance for all emotion primitives (activation, valence, and dominance) on IEMOCAP. Additionally, on the MSP- Podcast dataset, our method obtained considerable performance improvements compared to baselines.