NCSA Student Spotlight: Daniel George

By Boswell Hutson

For the past three years, Daniel George has been working towards a P.h.D in Astronomy at the University of Illinois at Urbana-Champaign, but when he started this journey in 2015, it was unlikely that he could have predicted the meteoric rise that was about to occur.

Since arriving on campus in August of 2015, George joined the Gravity Group at the National Center for Supercomputing Applications (NCSA) as a research assistant and immediately jumped into one of the most cutting-edge fields of astronomical study: gravitational waves – the same subject as the research that won the 2017 Nobel Prize in Physics. What’s more, George’s worked used a combination of artificial intelligence (AI) and machine learning (ML) to make the identification of these astronomical signals magnitudes faster, and use fewer computational resources.

“My research is using models from simulations, like that of black hole mergers, and developing an algorithm that can look at these models and see if it can identify a signal that is repeated in the actual LIGO data,” George explained.

“We’ve shown that we can detect the signals that have been detected so far, but much faster, and even in real-time, using this method. It also uses much fewer computational resources, like one CPU instead of the thousands of CPUs needed for the current methods.”

The real-world applications of George’s AI algorithms are evident, and the time and energy that they will save are unprecedented in astronomy. As such, George earned third place at the Grand Finals of this year’s ACM (Association of Computing Machinery) Student Research Competition.

This competition, one of the most fiercely competitive in the country, consisted of nearly 1,000 participants from 24 different conferences and required a peer-reviewed paper submission, a poster presentation and an oral presentation. The winners of each conference were then selected to attend the Grand Finals, where George took home third place in the entire country.

But participating in the ACM Grand Finals was only the beginning of George’s summer. He also spent time working as an NVIDIA Fellow and a Google X intern on their campus in Mountain View, California.

“I outlined my plans for my thesis and how I plan to develop these AI algorithms for gravitational wave signal processing,” George explained. “It’s interesting to NVIDIA because we’re using a lot of GPUs to train these neural networks.”

In return, NVIDIA provided George with funding to refine his algorithms. This experience prepared him for his time spent at Google X, one of the most novel working environments in the tech industry.

“Google X is this division in Google that works on sci-fi projects like self-driving cars, for example,” said George.

“Projects start in Google X, and it acts as an incubator. They have a diverse team of people across disciplines, essentially everything you would need to start a company. They have state-of-the art hardware right next to software engineers and MBA interns, all in the same building. They nurture these projects until they’re ready to spin out into their own company, and then they become their own company.”

While many of George’s projects at Google X are not publicly-announced, he offered a small glimpse into how his work can be applied to developing new products, a stark change from the academic world.

“I worked on deep learning and machine learning using neural networks and then applied that to problems similar to analyzing the LIGO data,” George said. “I worked on three projects, one was related to Project Loon, Google’s project to bring internet to rural areas via stratospheric balloons, and the two other projects were early-stage machine learning for time-series data.”

Later this month, George will defend his Ph.D. thesis and has already accepted an offer to return to Google X as a research scientist this year, showing how the skills and expertise learned at NCSA can prepare students for a plethora of careers, whether that be in research, academia or private industry.

Disclaimer: Due to changes in website systems, we've adjusted archived content to fit the present-day site and the articles will not appear in their original published format. Formatting, header information, photographs and other illustrations are not available in archived articles.