Memristor-Based Feature Recall Neural Network Circuit With Temporal Differentiation of Emotion and its Application in Parts Inspection

Abstract

From the perspective of time, the generation of emotion has not only emotional generalization but also emotional differentiation. Most studies of emotional systems only consider the present tense, and the past and future tenses are not considered. In this article, based on the bionic multiloop emotion learning model, a memristor-based feature recall neural network circuit with temporal differentiation of emotion is proposed. The circuit is composed of thalamus module, insular cortex module, anterior cingulate cortex module, sensory cortex module, amygdala module, and feature recall module. Feature recall module and emotion learning circuit are used to combine feature associative memory with emotion generation. Temporal differentiation of emotions is realized. The multiloop affective learning circuit also takes into account the activation patterns of different brain regions in response to positive and negative stimuli. The feasibility of the above circuits is verified by PSpice simulation software. The feature recall neural network circuit with temporal differentiation of emotion provides further reference for the bionic robot to realize inference prediction function.