Fusion Approaches for Emotion Recognition from Speech Using Acoustic and Text-Based Features

Different fusion approaches for emotion recognition from speech using acoustic and text-based features show improved performance with BERT embeddings compared to Glove embeddings.