Iolanda Leite’s research is at the intersection of Human-Robot Interaction and Artificial Intelligence. She takes inspiration from human theories of emotion and social relations to develop computational models that support the dynamics of social interactions between people and machines. These models rely on novel algorithms and machine learning techniques that allow robots to adapt their behavior to the particular context, user needs and preferences over time. Her long-term goal is to build autonomous social robots that can capture, learn from and respond appropriately to the subtle dynamics of real-world situations, allowing for truly useful and efficient long-term interactions with people.
What have I achieved during my fellowship?
We investigated whether social robots can positively influence group dynamics among children, specifically addressing intergroup bias. Our results indicate that robots can help the process of inclusion by mediating a group-based activity. The robot succeeded in encouraging outgroup children to act more outgoing and in increasing collaboration among ingroup and outgroup children. Further, children showed a higher level of prosociality after interacting with the robot.
To address these questions, we had to develop novel algorithms to enable robots to perceive the current group dynamics and chose how to assist such dynamics in real time. For example, we developed algorithms (based on heuristics and machine learning) so that robots can use their eye gaze to balance the level of participation of a group with different skill levels. These algorithms were evaluated in a human-robot language skill-dependent game, played by a native speaker and a second language learner, having the robot as the mediator.
Parts of this research were conducted in collaboration with another Jacobs Research Fellow (Wouter van den Bos). Our work was recognized with the Best Paper Award at the International Conference on Human-Robot Interaction, the premium venue for publishing Human-Robot Interaction research.
My plans for the future
We plan to further investigate how socially intelligent robots can enhance social interactions among children in other domains. For example, a postdoc in my team is studying how robots can encourage children to perform (repetitive) physical rehabilitation activities. Some of our experiments were conducted in the lab with adults due to limitations of perception technology (e.g., speech recognition and computer vision), which are still quite challenging especially in group interactions. Our immediate plans include addressing these technical limitations to determine whether the results found in the lab can extend to real world environments with children.