Engineering Normative and Cognitive Agents with Emotions and Values

While Artificial Intelligence (AI) has become part of our daily lives, there are emerging expectations for these AI systems to (1) reason over human factors with humans-in-the-loop (2) adapt to the changing environment or requirement in the real world. To achieve this goal, an agent must first incorporate emotions in its decisionmaking. Further, the agent must be capable of interpreting normative information from expressed emotions. Specifically, expressed emotions as information enable inference of non-observable mental states [16]. Furthermore, the agent must consider human values and preferences. The research presented here proposes an agent architecture to accommodate these factors.