Robots with emotional intelligence

ABSTRACT

This keynote talk will illustrate a basic set of skills of emotional intelligence, how they are important for robots and agents that interact with people, and how our research at MIT addresses part of the problem of giving robots such skills. One of the most important skills is the ability to perceive and understand expressions of emotion, which I will highlight by demonstrating our latest technologies developed to read joint facial-head movements in real-time and associate these with complex affective-cognitive states, and technologies to read paralinguistic vocal cues from speech. The latter have been made open-source and are available for free. I will also show some non-traditional ways robots might sense and learn about human emotion, and ways they can respond to what they sense that can help or hurt people. I will discuss social and ethical issues these technologies raise. Finally, I will present some new possibilities for robots to both learn from people and help teach skills of emotional intelligence to people, especially to those with nonverbal learning impairments who may want to learn these skills, including many people with diagnoses of autism spectrum disorders such as Aspergers Syndrome.

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
      I.2.9 Robotics

General Terms:
Human Factors, Performance, Reliability, Security

Keywords:
affective computing, autism, deception detection, emotion recognition, emotional intelligence, empathic technology, facial expression recognition, physiological sensing, prosody analysis, system