Toward Machines with Emotional Intelligence

M.I.T. MediaLaboratory

Talk:  Friday, April 9, 2004 
Talk: 10:00-11:00 a.m<
Location: Room 3105 Engineering Bldg.

Host: A. Jain

Abstract: Over 70 studies on human-machine interaction in the last decade have pointed to an intriguing phenomenon: People tend to interact with machines in a way that is very similar to how they interact with each other, even when the machine is not a robot, agent, or other kind of obviously social actor.  This finding holds true even for intelligent computer science and engineering students who know that machines don't have feelings.  The finding suggests that many of the more subtle skills critical for human-human interaction are also significant for human-computer interaction.

The skills of "emotional intelligence" have been argued to be among the most important for people, even more important than mathematical and verbal intelligences.  Emotional intelligence includes the ability to recognize emotion -- to see if you're irritated or annoyed someone, pleased or displeased them, bored or interested them.  It includes the ability to know when to show emotion (or not), and how you should respond to another's emotions, as well as many other skills.

In this talk, I'll describe how we're giving computers new skills of intelligence, specifically the ability to recognize and respond appropriately to human emotion.  I'll show examples of systems that try to assess interest, frustration, stress, and a range of other states that occur when interacting with computers.  These systems involve new kinds of sensing for desktop, wearable, and other environmental interfaces, as well as the development of new signal processing, pattern recognition, and machine learning algorithms for drawing inferences about the multimodal data.

Current applications include human learning, usability feedback, health behavior change, and human-robot interaction.

Biography: Rosalind W. Picard is founder and director of the Affective Computing Research Group at the Massachusetts Institute of Technology (MIT) Media Laboratory and is co-director of the Things That Think Consortium. In 1984, she earned a Bachelor in Electrical Engineering with highest honors from the Georgia Institute of Technology. Picard earned Master and Doctorate degrees, both in Electrical Engineering and Computer Science, from the Massachusetts Institute of Technology (MIT) in 1986 and 1991, respectively. The author of over a hundred peer-reviewed scientific articles in pattern recognition, multidimensional signal modeling, computer vision, and human-computer interaction, Picard is known internationally for pioneering research into digital libraries and content-based video retrieval. She is co-recipient with Tom Minka of a "best paper" prize (1998) from the Pattern Recognition Society for their work on interactive machine learning with multiple models. She has served two terms as Associate Editor of IEEE Trans. on Pattern Analysis and Machine Intelligence, and is active on several scientific program committees and review boards. Her award-winning book, Affective Computing, (MIT Press, 1997) lays the groundwork for giving machines the skills of emotional intelligence.