In this study, we compared three architectures for the task of age and gender recognition from voice data: Long Short-Term Memory networks (LSTM), Hybrid of Convolutional Neural Networks and ...