Carnegie Mellon University
Two computer science students at Carnegie Mellon University in Qatar were recognized for their excellence in pioneering innovative scientific research in computing at the 2011 Qatar Foundation Annual Research Forum. Amna Al Zeyara and Nawal Behih, both sophomores, received the Best Student Research Award in Computing and a 20,000 Qatari Riyal grant for their work on enhancing facial expressions in bilingual humanoid robots. The testbed for the research project was Hala, a bilingual multi-cultural robot platform developed by Carnegie Mellon.
“We are very proud of both Amna and Nawal for this significant achievement,” said Majd Sakr, Ph.D., assistant dean for research at Carnegie Mellon Qatar. “They started working on this research project as freshman computer science students. They managed to quickly learn the scientific methodology of research, how to evaluate state-of-the-art techniques in robotics and build upon what they have learned to improve the behavior of humanoid robots.” Sakr and Imran Fanaswala, a research programmer at Carnegie Mellon Qatar, mentored the team.
Hala is a 3-D animated female robot that interacts with visitors, switching between English and Arabic languages and crafting responses that are culturally sensitive to the user. To do this, Hala takes linguistic cues from the user to determine the cultural background – ensuring effective communication and minimizing misunderstandings. In the future, Carnegie Mellon sees robots like Hala deployed as intermediaries between cultures, closing the communication gap between parties with different backgrounds.
Hala is currently undergoing major enhancements to help improve her embodiment of the Arabic culture. Amna and Nawal’s project addressed existing issues with animating natural facial expressions and implementing more realistic lip-movements (“visemes”) to synchronize with the Arabic language. Natural expressions help engage the users and improve their experience by enhancing non-verbal aspects of communication.
“We used existing research and observed each other to develop visemes that accurately capture Arabic pronunciations,” Amna said. “We also authored an in-house tool which allows non-programmers, such as artists, to manipulate the face in real-time to create natural expressions.”
As the result of the students’ work, Hala now has 11 new facial expressions for a more natural looking and behaving robot. The work has also pioneered the first implemented subset of Arabic visemes on a robot.
Carnegie Mellon University in Qatar engages students at all level to engage in active, regionally relevant and impactful research. Through this process, the students learn new skills that help shape their future and career.