The Department of Electrical and Computer Engineering welcomes new faculty member

Almost anyone who has ever streamed a film can relate to the twinge of frustration that occurs when a pivotal scene is marred by the sudden appearance of blocky patterns or jagged outlines. Those distortions, called digital artifacts, are the result of image compression — a technique applied to a graphics file to reduce its size. A compressed image takes less bandwidth to transmit or download, resulting in less network traffic, lower energy usage, and faster content delivery — all important considerations at a time when well over half of all internet content is composed of videos and images.

Associate Professor Jona Ballé, the newest member of Tandon’s Department of Electrical and Computer Engineering, has set out to improve the situation. “Image compression is functional, but it doesn’t take into account how human beings see,” they explain.

Ballé, who grew up in Cologne, Germany, was educated at RWTH Aachen University, one of the country’s top engineering schools, where their doctoral studies focused on image processing and understanding human perception. Impressed by the research being done at NYU, they subsequently traveled to Washington Square to do post-doctoral work at the Center for Neural Science’s Laboratory for Computational Vision.

When they arrived, in early 2013, researchers working to create better image compression algorithms had not yet harnessed the power of machine learning, but within a few years, Ballé had helped change that. “This was the tool I had been missing,” they recall. “By 2017, a new paradigm in data compression was being ushered in, and machine learning was allowing us to develop algorithms for better modalities and error metrics.”

Ballé made a move that year to the West Coast to join a team at Google that was focused on the same questions they were interested in: how could machine learning be harnessed to build better data compression algorithms, and make the results look more realistic?

While they enjoyed working there, Ballé eventually missed both the East Coast and academia, where their research could return to tackle these questions in a more fundamental way. When the opportunity arose to join the faculty at NYU Tandon, it felt like a homecoming of sorts. “I’ll be teaching graduate students, and we’ll be taking very empirical, hands-on approaches,” they say. “I’m already getting a lot of interest from potential Ph.D. candidates and post-docs, so it’s very exciting.”

Ballé, who is non-binary, also welcomes the chance to be a presence for others who might not ever have seen someone who looks like them or shares similar experiences in a STEM setting. “I wouldn’t use the term ‘role model,’ because that might be too lofty a way to describe it,” they say. “But it’s important to be visible – to be seen and acknowledged – so everyone realizes that STEM can be for them.”