Academics

IST junior spends two summers at IBM Research, files patent for prediction tool

Yuya Ong (left), a junior in the College of Information Sciences and Technology, spent two summers as an intern with IBM Research. During his stay in San Jose, California, this past summer, he was able to meet Pieter Abbeel (right), a well-known machine-learning researcher at UC Berkeley, at a workshop. Credit: ProvidedAll Rights Reserved.

UNIVERSITY PARK, Pa. — Yuya Ong, a junior in the College of Information Sciences and Technology, feels fortunate to be at the forefront of the rapidly-growing data sciences field.

“Every day when I wake up, I realize that there are new opportunities out there and new problems to be solved,” he said. “Waking up to that is really exciting.”

Ong explored those emerging opportunities through his summer internship with IBM Almaden Research Center in San Jose, California, where he focused on researching a new architecture for predictive analytics using deep learning methods. Specifically, he was tasked with developing a new algorithm for time series forecasting, in which a series of past values, such as stock trends or temperatures, is used to predict the next set of values.

“When [IBM Research] defined the problem, it was given to me in an open ended way,” said Ong. “It was up to me to see how I can frame the problem and devise a solution.”

He drew on the knowledge he gained in the College of IST and specifically in his Applied Data Sciences course, taught by James Wang, professor of IST, and on the guidance of his IBM Research mentor, Mu Qiao, a Penn State doctoral graduate and former student of Wang’s.

Ong recently applied for a patent for a the algorithm he developed during his internship. The algorithm borrows techniques from computer vision to transform the data into a structure where it can learn novel features from the data. The algorithm could enhance multivariate time series forecasting to the benefit of industries such as stock market analysis, Internet of Things and Cloud infrastructure, weather forecasting, and health analytics. The preliminary testing results of the algorithm show significant improvements over state-of-the-art deep learning methods and comparative results over classical methods.

“When I presented my work at the poster presentation, researchers from across various disciplines such as physics, biology and chemistry came up to me and said `We want to make use of this,’” said Ong. “I am really excited that my work can be applied in different industries. That is big.”

“A lot of the classes in the College of IST, especially the data sciences courses, reflected what the industry needs were,” he added. “In data visualization, some of the theory provided a great framework for designing and building visualizations and tools to understand how to better build interpretable Deep Learning models.”

Through his internship, he also worked on developing the modeling component for IBM's latest verifier system, Crypto Anchor Verifier. This new technology brings utilizes AI and optical imaging to help prove the identity and authenticity of objects, such as diamonds.

This isn’t the first time that Ong has worked for IBM Research, as he also completed an internship with the company at the end of his freshman year. It was during that time that Ong created a solution to identify whether a document an individual is about to upload contains private or sensitive information.

“One of the things that we innovated on was the use of contextual text features [rather] than the use of key-words or dictionary-based approaches. We used a deep learning approach to pick up those subtle nuances to tell whether something is sensitive or not, at various scales in real-time,” Ong explained.

The project piqued Ong’s interest in data privacy and security, and in the two years that followed, he took several classes on those topics. Meanwhile, IBM launched new data privacy and security processes to ensure new products align with the recently-passed General Data Protection Regulation (GDPR).

Thanks to the evolving data sciences curriculum at Penn State, Ong was able to stay current on these trends to position him for his junior year internship with IBM Research.

“When the industry changes, it shifts the culture as well as the way the company operates,” Ong said. “Witnessing such observations, in such a short period of time, was interesting to see how the company reacts to these current events.”

Now, with two summers at IBM Research under his belt, Ong advises all College of IST students to take advantage at the internship and career development opportunities available to them.

“I strongly believe in having hands-on experience,” he said. “It’s not enough just to pick up books and attend classes. You’re doing this for a purpose.”

He recalled some words he’d once heard regarding the countless open and research problems that exist in the world.

“Seek opportunities where you can apply your skill and tackle those challenges,” he advised. “Make a big difference.”

Last Updated September 7, 2018