In this paper, we are interested in exploiting textual and acoustic data of
an utterance for the speech emotion classification task. The baseline approach
models the information from audio and text independently using two deep neural
networks (DNNs). The outputs from both the DNNs are