Socio-Cognitive Interaction Between Human and Computer/Robot for HCI 3.0

摘要

With an increasing number of AI applications in everyday life, the discussion on how to improve the usability of AI or robot devices has been taking place. This is particularly true for Socially Assistive Robots (SAR). The presented study in this paper is an attempt to identify speech characteristics such as sound and lexical features among different personality groups, specifically per personality dimension of Myers-Briggs Type Indicator (MBTI). and to design an Artificial Neural Network reflecting the correlations found between speech characteristics and personality traits. The current study primarily reports the relationship between various speech characteristics (both mechanical and lexical) and personality dimensions identified in MBTI. Based on significant findings, an ANN (Artificial Neural Network) has been designed in an effort to predict personality traits only from speech processing. The current model yielded 75% accuracy in its predictive ability, warranting further attention to the applicability of speech data in developing and improving various domains of human-computer interactions.