Advancements in Voice Assistants: A Study of Speech Recognition and Emotional Intelligence

Today the interaction with a user is achieved in a more advanced way due to the advanced systems, and researchers are coming up with innovative solutions to improve these systems. This paper suggests an automatic Speech Emotion Recognition (SER) algorithm that employs deep learning techniques which combine three-dimensional convolution and recurrent neural networks, and it operates in real time. Taking advantage of advanced feature extraction methods, such as Mel-Frequency Cepstral Coefficients (MFCCs)and chroma features, the model targets happy, sad and angry moods. We show that our system has high accuracy and good response for different emotional states. The proposed research adds to the increasing literature in the affective computing area by focusing on SER system development as well as providing a model for the integration of emotional intelligence into AI powered voice assistant systems.