AP NEWS

2 CMU researchers to work part-time at Facebook AI lab in Pittsburgh

May 10, 2018 GMT

Facebook appears to be eyeing robotics and artificial intelligence with two Carnegie Mellon University researchers who will work at its new AI lab in Pittsburgh.

Jessica Hodgins and Abhinav Gupta, both professors of robotics and computer science at CMU, will split time working for Facebook and continue teaching and research projects at CMU.

Hodgins and Gupta declined to comment on their work with Facebook and referred questions to the company. A Facebook spokesman said the company will announce more details about the Pittsburgh lab and the scope of its work in the coming weeks.

Facebook will also open an AI lab in Seattle and has hired Luke Zettlemoyer, a professor at the University of Washington and researcher at the Allen Institute for Artificial Intelligence.

Yann LeCun, director of Facebook AI Research, made it clear that Facebook was not poaching talent from the universities. Facebook has similar AI research labs in New York, Paris, Montreal, Tel Aviv and Menlo Park, Calif. Each lab has relationships with nearby universities.

“Facebook is careful not to deplete universities from their best faculty, by making it easy to maintain sizeable research and teaching activities in their academic labs,” LeCun wrote on his Facebook page. “Unlike others, we work with universities to find suitable arrangements and do not hire away large numbers of faculty into full-time positions bottled up behind a wall of non-disclosure agreements. We contribute to local ecosystem.”

Uber famously lured about 40 researchers from CMU when it started its self-driving car project in Pittsburgh. JPMorgan last week hired Manuela Veloso, who will take a leave of absence from her post as head of CMU’s Machine Learning Department.

Hodgins is the former director of Disney’s research lab at CMU, which closed in February. Her research has focused on making robots and their interactions more human-like.

Gupta studies computer vision and perception and how robots interact with objects around them. He worked with a robot that taught itself how to grasp objects.

It is unclear what role robotics will play for Facebook. Perception, computer vision and human-computer interaction could help Facebook better manage content and improve how people use the platform.

Facebook operates a research and development facility for Oculus, which it bought for $2 billion in 2014. The virtual reality lab at Schenley Place in Oakland has kept secret what it is working on and has not allowed tours or access.

Facebook is betting on artificial intelligence to solve most of its most difficult problems, from detecting fake accounts and Russian bots to quashing bullying and harassment to taking down inappropriate content. CEO Mark Zuckerberg mentioned AI more than 30 times during his testimony on Capitol Hill last month, according to the Washington Post. He said the AI utopia is still five to 10 years away.

Facebook will send Jerome Pesenti, vice president of AI, to the White House on Thursday for a summit on artificial intelligence. Executives from nearly 40 other companies, including Amazon, Google, Intel and Ford, will be there. CMU President Farnam Jahanian and Andrew Moore, dean of CMU’s School of Computer Science, also plan to attend.

Facebook is already using artificial intelligence to take down bad content such as posts containing terrorist videos, hate speech, pornography and violence before it is reported, Guy Rosen, vice president of product management for Facebook wrote in a blog post last week.

“It’s taken time to develop this software — and we’re constantly pushing to improve it,” Rosen wrote.

Aaron Aupperlee is a Tribune-Review staff writer. Reach him at aaupperlee@tribweb.com, 412-336-8448 or via Twitter @tinynotebook.