Artificial Psychology Revisited : Constructs for Modeling Artificial Emotions

Understanding the DNA of feelings and emotions are becoming increasingly important mechanisms for facilitating learning in intelligent systems. Here we present constructs for modeling human emotions within a foundation of artificial cognitive architectures[1] We model artificial neural emotions through the use of a weighted spatial-temporal emotional memory system, based upon knowledge relativity threads [2&3] human Autonomic Nervous System States. We artificially evolve the granularity of “emotional triggers” to determine importance thresholds which are context specific and relative to sensory inputs and environmental conditions to advance the capability of emotional learning and processing. We believe this has the potential to enhance artificial neural processing environments by allowing emotional memories and emotional learning to facilitate coalitions and cooperation between artificial neural intelligent software agents. We believe shared emotional states between intelligent software agents will more easily allow information sharing between agents, based on the “emotional reaction” to the systems sensory inputs, much in the way humans do.