Two professors will delve into the privacy issues resulting from the data-based training technique known as "deep learning," courtesy of a grant from Google.
Pennsylvania State University's Adam Smith, an associate professor of computer science and engineering, and Vitaly Shmatikov, an associate professor of computer science and a visiting faculty member at Cornell Tech, will investigate the chinks in the armor of deep learning — that is, using algorithms to model high-level abstractions in data.
In particular, Smith and Shmatikov will focus on how private information can be leaked from the massive data sets employed by machines using deep learning, according to a news release from PSU.
“Deep learning is already widely used, especially at Google, to recognize speech and images and to do lots of other things, like drive cars,” Smith said in the news release. “Deep learning is often applied to very sensitive data, however. We are interested in making sure that sensitive, private data remain private.”
The grant is one of hundreds of Google Faculty Research Awards the company distributes each year to fund research being conducted in computer science, computer engineering and similar fields.
Smith and Shmatikov’s work includes “the development of a system for privacy-preserving deep learning,” according to the release.
Google stands to benefit from the research, as both professors are encouraged to speak to the tech giant’s employees about their work. But others will get an education as well, once the findings are made public.
The study of data science is a booming field in higher education. Schools such as Michigan State University have expanded their analytics expertise by bringing on a new kind of professor, one specializing in Big Data. Enrollment for courses in computer science often goes beyond capacity due to interest. The Bureau of Labor Statistics projects a massive shortage in the IT workforce by 2020: There will be 1.4 million openings, but only 400,000 computer science graduates with the necessary skills to fill the positions.