The Grounding Problem of Modeling Emotions in Adaptive Artifacts

1
The Grounding Problem of Modeling Emotions in
Adaptive Artifacts

• Thomas Wehrle
• GIL,University of Geneva,Switzerland
• by
• Onur Ozyer
• oozyer_at_cs.ucf.edu

2
Motivation

• Emotions are important for mammals to
• reach the multiple goals with limited resources.
• Intelligent agents need emotions to be
• embedded in for this reason. But both AI and
• synthetic emotions have difficulty to describe
• their counterpart. This is because of the
• misconceptions in the field and multiplicity
• of the phenomena.

3
Overview

• A- Motives for emotional agents.
• B- Description of two design approaches for
  emotion
• modeling, and their relations with motives.
• C- Adaptation in affective systems and grounding
• emotions.

4
Motives

• Science Learn much of the emotions and an their
• mechanisms by building a model of a theory with
  agents, and
• compare the results with the predictions.
• Engineering Design and implement useful
  artifacts for concrete
• tasks with a good performance value.
• Human Computer Interaction Build systems with
  better HCI
• in terms of usability, performance and
  acceptability.

5
Modeling Approaches

• a) Black Box Modeling Gives output similar to
  naturally
• occurring systems with out taking the underlying
  processes into
• consideration. These models focus on input-output
  relationship and
• they are practical for decision-making.
• b) Process Modeling Simulates the underlying
  mechanisms
• generating the emotions in a degree. Effects of
  inputs on the sub-
• mechanisms and inter-relations between these
  mechanisms should
• be well known.

6
Appraisal Theories

• Appraisal Theory Elicitation of an emotion
  process and its
• differentiation into different categorization
• a) Componential Approach
• 1) Emotions are elicited by a cognitive
  evaluation of antecedent
• events.
• 2) Patterning of reactions are shaped by this
  appraisal process.
• Appraisal dimensions are used to evaluate
  stimulus, in an adaptive
• way to the changes.
• Appraisal Dimensions Evaluation of significance
  of event, coping
• potential, and compatibility with the social
  norms.

7
Appraisal Theories

• b) Component Process Model Has less evaluation
  criteria and they
• are processes sequentially. Stimulus evaluation
  check (SEC) is built
• on this idea. In emotional states, the
  components of the
• emotion synchronize their functioning
  temporarily, then stimulus is
• evaluated in the system, then control is returned
  to independent
• sub-systems. During evaluation, all systems are
  highly related and
• multi feedback and multi feedforward operations
  occur.

8
Adaptation

• Emotions must be adaptive since they are results
  of interaction
• and synchronization of complex subsystems, as a
  response to
• changes in inner and external states. Individual
  adaptability can be
• implemented in three levels,
• a) Level of single values, changing the
  parameters of the
• stimulus evaluation systems.
• b) Level of emotion specific value patterns,
  change in the
• significance of different appraisal systems.
• c) Level of action readiness, modification action
  repertory to an
• emotional state.

9
Conclusion

• There are different underlying mechanisms of the
  emotions, the
• question is to what extent the complexity of
  these mechanisms
• should be abstracted, or put into the black box.
  Grounding
• emotions examines this abstraction.
• Introduce a value system to the robot, allow the
  robot establish its own emotional categorization.
  These categories can be different than our
  emotions.
• Put human emotions categories into agents. But we
  cant ground them properties, tasks and
  environment of agent is different.