Ladder Networks for Emotion Recognition: Using Unsupervised Auxiliary Tasks to Improve Predictions of Emotional Attributes

TL;DR

This paper introduces ladder networks with unsupervised auxiliary tasks to enhance emotion recognition by jointly learning emotional attributes and reconstructing hidden representations, improving performance especially in semi-supervised scenarios.

Contribution

It proposes a novel regularization approach using ladder networks with denoising auxiliary tasks for better emotion attribute prediction.

Findings

Ladder networks outperform baseline models in emotion recognition tasks.

Unsupervised auxiliary tasks improve feature representations.

Potential for semi-supervised emotion recognition with limited labeled data.

Abstract

Recognizing emotions using few attribute dimensions such as arousal, valence and dominance provides the flexibility to effectively represent complex range of emotional behaviors. Conventional methods to learn these emotional descriptors primarily focus on separate models to recognize each of these attributes. Recent work has shown that learning these attributes together regularizes the models, leading to better feature representations. This study explores new forms of regularization by adding unsupervised auxiliary tasks to reconstruct hidden layer representations. This auxiliary task requires the denoising of hidden representations at every layer of an auto-encoder. The framework relies on ladder networks that utilize skip connections between encoder and decoder layers to learn powerful representations of emotional dimensions. The results show that ladder networks improve the performance of the system compared to baselines that individually learn each attribute, and conventional denoising autoencoders. Furthermore, the unsupervised auxiliary tasks have promising potential to be used in a semi-supervised setting, where few labeled sentences are available.