Semisupervised Autoencoders for Speech Emotion Recognition

Abstract

Despite the widespread use of supervised learning methods for speech emotion recognition, they are severely restricted due to the lack of sufficient amount of labelled speech data for the training. Considering the wide availability of unlabelled speech data, therefore, this paper proposes semisupervised autoencoders to improve speech emotion recognition. The aim is to reap the benefit from the combination of the labelled data and unlabelled data. The proposed model extends a popular unsupervised autoencoder by carefully adjoining a supervised learning objective. We extensively evaluate the proposed model on the INTERSPEECH 2009 Emotion Challenge database and other four public databases in different scenarios. Experimental results demonstrate that the proposed model achieves state-of-the-art performance with a very small number of labelled data on the challenge task and other tasks, and significantly outperforms other alternative methods.

Adversarial Training in Affective Computing and Sentiment Analysis: Recent Advances and Perspectives [Review Article]

“…Similarly, Cummins et al [43] utilised CNNs pre-trained on large amounts of image data to extract robust feature representations for speech-based emotion recognition. More recently, neural-network-based semi-supervised learning has been introduced to leverage large-scale unlabelled data [13], [44].…”

Autonomous Emotion Learning in Speech: A View of Zero-Shot Speech Emotion Recognition

“…In the attribute-learning phase, we consider three approaches for fi(•)s, namely, shallow-structure Multi-Layer Perceptron (MLP) networks [27], Support Vector Regression (SVR) [28], and Ridge Regression (RR) [15]. The MLP consists of a two-hidden-layer structure, with 12 selections of the hiddenlayer neurons as: (32,8), (32,16), (64, 16), . .…”

“…Prominent directions concerning SER have focused on diverse topics such as data collection [3], data enrichment [4], deep learning [5], feature enhancement [6], and transfer learning [7]. Nevertheless, most current research on SER is, arguably, focused on the passive learning of emotional states, which need fully, or at least partially labelled training samples to learn reasonable models [8]. Furthermore, passive approaches are unsuitable for labelling samples which do not have an adequate amount of matched training data.…”

Speech emotion classification using fractal dimension-based features

“…In this way, the neural network becomes a "black-box" analysis technique for speech emotion feature extraction. Nevertheless, acoustic (or any other) analysis of the speech signal still remains relevant topic in speech emotion recognition [13,28,55].…”