UNIVERSITY PARK, Pa. — Emboldened by his education, Jianbo Ye, a doctoral student in the College of Information Sciences and Technology (IST), will soon help Amazon use technology to revolutionize the way we live.
After he graduates in the spring, he will begin working as a researcher at Amazon Lab126, the secretive research and development unit located in the Silicon Valley. Among its many innovations, the lab is responsible for the Alexa and other intelligent consumer products.
“I’m excited to join them because these are projects people haven’t done before,” Ye said. “It’s going to be a challenging venture.”
James Wang, professor of IST and one of Ye’s advisers, is confident Ye’s research explorations have prepared him well for the challenge.
Wang said that his position at Amazon will allow Ye “to work side-by-side with some of the world's prime information technology talents.”
During his time at IST, Ye made several fundamental and innovative contributions to his field through research. In his work under an NSF-funded project, jointly conducted by IST and researchers in the departments of psychology and statistics, Ye authored a paper that proposed a new method to evaluate and improve the quality of crowdsourced affective data. In Wang’s exploration of coding human emotions, this work is vital.
“The data we are collecting, utilizing Ye’s work, will be highly valuable in developing future robots and computers that can be empowered with emotional intelligence to better interact with people,” Wang said.
However, Ye considers his most groundbreaking work to be in the field of machine learning, working primarily with his co-adviser Jia Li, professor of statistics.
Machine learning, which is a branch of artificial intelligence to program a computer to “learn” and improve its own functions and approaches to problems, is a rapidly growing area of research.
For his dissertation, Ye proposes a new approach to this concept using optimal transport.
“It’s a new actively pursued direction for machine learning,” he said. “We are devoted to doing research in this area.”
Optimal transport, in the context of machine learning, incorporates the element of prior knowledge to help make sense of large amounts of data. For example, if a program was tasked to gather all the social media conversation about a recent snowfall, conventional methods would compile sentences only containing the word “snow.”
Ye proposes coding an existing database of knowledge that would allow the computer to make connections between similar words like “flurries” or “flakes” to create a more comprehensive understanding of the data.
“When you develop your knowledge on a topic, you are able to comprehend more complicated scenarios,” he said. “It’s similar in machine learning. If you have rich prior knowledge on a topic, you can apply it to comprehend data.”
Ye’s paper on the topic, “Determining Gains Acquired from Word Embedding Quantitatively using Discrete Distribution Clustering,” was presented at the annual meeting of the Association of Computational Linguistics. He hopes this approach will help refine document analysis so computers will be able to sort and derive more impactful meaning from large collections of text.
Through his dozen publications, it’s no doubt these impressive contributions led to his job offer from Amazon.
“He will leverage the knowledge and skills gained at our Ph.D. program [to] futuristic device inventions that will benefit millions of people in their daily lives,” Wang said.
The position is a perfect fit for Ye, who is looking forward to a fulfilling career in industry.
“I want to make an impact in the real world,” Ye said.