AI and HRI
I work on Artificial Intelligence (AI) algorithms that improve the real-time perception, reasoning and expressive capabilities of embodied social agents. I also plan and conduct controlled and in-the-wild Human-Robot Interaction (HRI) studies mostly in assistive, education and entertainment domains.
Worked on a fast-paced keyword spotter for children aged 5-10. Then, created the artificial intelligence for a robot co-player that was initially unengaging but became engaging when we adapted its prosody to each child's speech. *Best technology paper award at HRI 2017
Social Presence Guidelines
Built a custom digital table and implemented an adaptation of the Risk game. Established a set of guidelines for developing socially present agents and developed the autonomous behavior for a robot that plays Risk against three human players. *Best paper award at ACE 2012
Worked on the problem of using semi-situated crowdworkers to automatically extend a robot's dialog knowledge and enable it to express believable and appropriate verbal and nonverbal behaviors. Also developed a module that matches RFID tags with users detected on a Kinect camera.
I have extensively worked on the topic of closing the loop from affective recognition to generating empathic behavior in robots.
After the “novelty effect” people usually lose interest and change their attitude towards robots. One of my main long-term research goals is to successfully achieve long-term relationships between robots and humans.