Prediction of Emotional Empathy in Intelligent Agents to Facilitate Precise Social Interaction

Abstract

The research area falls under the umbrella of affective computing and seeks to introduce intelligent agents by simulating emotions artificially and encouraging empathetic behavior in them, to foster emotional empathy in intelligent agents with the overarching objective of improving their autonomy. Raising the emotional empathy of intelligent agents to boost their autonomic behavior can increase their independence and adaptability in a socially dynamic context. As emotional intelligence is a subset of social intelligence, it is essential for successful social interaction and relationships. The purpose of this research is to develop an embedded method for analyzing empathic behavior in a socially dynamic situation. A model is proposed for inducing emotional intelligence through a deep learning technique, employing multimodal emotional cues, and triggering appropriate empathetic responses as output. There are 18 categories of emotional behavior, and each one is strongly influenced by multimodal cues such as voice, facial, and other sensory inputs. Due to the changing social context, it is difficult to classify emotional behavior and make predictions based on modest changes in multimodal cues. Robust approaches must be used to be sensitive to these minor changes. Because a one-dimensional convolutional neural network takes advantage of feature localization to minimize the parameters, it is more efficient in this exploration. The study’s findings indicate that the proposed method outperforms other popular ML approaches with a maximum accuracy level of 98.98 percent when compared to currently used methods.