Social & Emotional Agents

Actual lively interaction between humans is based on active communication using language, non-verbal gestures and emotions. Currently artificial companions are capable of using language as a mean of communication with humans while emotions are not well exhibited by artificial agents. Affective communication and empathy is highly recommended for realistic and long interactions. LIREC researchers focus on implementing affective and emotional agents using certain frameworks and systems to model emotions and their relations to behaviors and actions.

Empathy

Empathy is a feeling evoked as a result of a person attending and understanding the emotional expression of another person. It is one of the methods used by humans to maintain a relationship Bickmore & Picard (2003). Social agents that are humans-relations inspired could also use empathy as part of their communication and interaction methods. Virtual and social agents have already been acting empathetic with other agents and humans. As an example, the use of empathetic characters offers great advantages in children educational software applications. These interactive characters can autonomously act according to their own state and emotions resulting in unlimited possibilities of behaving causing the predictability of their actions to be minimized while the believability to be increased. Reaching such autonomous and empathetic characters requires the implementation of an emotional framework as Thomas and Johnston put it:” it has been the portrayal of emotions that has given the Disney characters the illusion of life” (Thomas, F., Johnston, O.: Disney Animation: The Illusion of Life. Abbeville Press, New York 1981).

Appraisal - Theory of Emotion and Cognition

Relating emotions with human behavior and actions is the essence of the appraisal theory. This idea goes back to the old time of Aristotle. It basically describes the appraisal process that is a cognitive process in which an individual generates emotions according to the events in the environment impacting its well being resulting in a subjective judgment on how bad or good an event is (Lazarus 1991). It is agreed upon that there are two types of appraisal. The first appraisal refers to whether the event is relevant to the agent or not. The secondary appraisal refers to coping options if the event is irrelevant to the agent and creates a threat. In order to deal with a threat to the agent’s goal, the agent can either try to plan for an action that will avoid the threat, or can simply accept that the goal will fail and lower its importance in order to avoid negative feelings about it. Personality affects coping strategies in agents similarly to humans where we all differ in our coping strategies against similar events. Certain aspects of this theory make it reliable and adequate for describing artificial agents emotions. Since the appraisal precedes emotions, it complies with the computational process that precedes eliciting the agent’s emotion to a certain situation. Differences in appraisal account for individual differences that will not restrict the computationally generated emotions to follow one pattern; rather there will be differences in emotional expressions of different agents.

OCC Theory

Many theories model emotions and related reasoning. One main and well-known model is the OCC theory (Ortony et al., 1988). Many researchers think of this theory as the ultimate solution to build emotions for synthetic character. It is widely used because the authors developed a computational model for emotions (Bartneck 2002) making it easy to implement in computer science because of its straightforward appraisal theory. OCC defines a group of emotions in 22 categories organized in a specific hierarchy. This model has five main phases to process an event starting with classifying the emotional categories for an event, object or action. The intensities of categories are quantified and used to define the emotional values that will be mapped and expressed by facial expressions or body movements or any other form possible for the agent.