IST researcher receives Amazon Research Award for work in affective computing

UNIVERSITY PARK, Pa. — James Wang, professor of information sciences and technology, has received a 2018 Amazon Research Award for his work in affective computing. His project, "An Open Emotion Understanding Dataset," will receive gift funding of $80,000 and a cloud computing credit of $20,000.

With the funding, Wang and his students will expand on previous research examining how to train a computer to recognize human emotional expressions from body movements. Wang’s team and their collaborators in the Penn State departments of Psychology and Statistics have already built a database of more than 13,000 individuals displaying nearly 10,000 body movements from movie clips and utilized human annotators to identify the emotion for each.

“We have been working on this research for two years,” said Wang. “It’s very exciting that Amazon sees the value in further expanding this effort.”

Prasenjit Mitra, associate dean for research at the College of IST, noted that the ability for computers to understand nonverbal cues will improve the quality of human-machine communication in the near future.

“Machine understanding of human emotions by observing a person via computer vision techniques is a first, necessary step toward creating machines that can interact better with human beings,” he said. “Dr. Wang’s pioneering work is a crucial first step toward achieving this vision.”

The Amazon Research Awards program offers awards of up to $80,000 to faculty members at academic institutions worldwide for research in areas from computer vision to online advertising to security, privacy and abuse prevention. Only 82 awardees, including Wang, were selected this year to receive funding. Wang is the first Penn State researcher to receive an Amazon Research Award.

“It shows that we are at the forefront of sciences and technology in the information domain,” said Wang. “I haven’t seen a lot of similar projects, perhaps because it is challenging and costly to collect high-quality affective dataset. This dataset we are developing will help reduce the barrier of entry for researchers interested in emotion research.”