Facebook AI partners with UC Berkeley to help accelerate advancements in AI

June 01, 2019

Facebook AI believes strongly in open collaborations with top academic institutions to help the entire AI community accelerate research in this field. To that end, Facebook AI is proud to support Berkeley Artificial Intelligence Research (BAIR) in establishing the BAIR Open Research Commons, a new program at UC Berkeley launched to help advance cutting-edge AI research between industry leaders and the university's students and faculty. In keeping with Facebook AI's commitment to open research, the Commons agreement promotes the sharing of new ideas through ongoing on-campus partnerships and collaborations with researchers at Facebook as well as Amazon, Google, Samsung, and Wave Computing. All research is expected to be published publicly, with code open sourced and accessible to the entire research community.

The creation of this partnership was greatly aided by dual-affiliated professors, including Jitendra Malik, who is both a Facebook AI research director and also a founding director of BAIR. Malik officially announced the partnership and Facebook's participation in early May, when he discussed the Commons' goal of advancing graduate student education while also accelerating AI research in an open and accessible way.

The partnerships will play an important role in helping the AI research community, Malik says, and it will also provide UC Berkeley students and faculty the opportunity to collaborate with industry leaders in the private sector. “The talent at BAIR is unparalleled and situated within one of the leading public research universities in the world,” he says.

Under the program's open-science framework, Facebook and BAIR plan to launch as many as eight joint research initiatives annually over the next five years, pairing UC Berkeley graduate students and their faculty advisors with Facebook's leading AI researchers, and then focusing on robotics, computer vision, reinforcement learning, computational linguistics, and other topics.

“We will work through triangles of partnerships, with one end being the BAIR student, one being a Facebook AI researcher, and the third vertex a UC Berkeley faculty member,” Malik says. “To form a relationship, we need this triplet. And what goes inside the triangle is a rigorous and impactful research project.”

Facebook's partnership with BAIR is managed by a Facebook AI scientific committee that is charged with identifying research projects with the potential to both advance graduate student research and also make an impact in the real world. Facebook AI research scientist and scientific committee chair Piotr Dollar says, “I believe the partnership with BAIR has awesome potential, and we plan to work closely with our faculty colleagues to realize the possibilities.”

Georgia Gkioxari, a Facebook AI research scientist who did her doctoral research at BAIR, agrees. “Given our closely aligned research agendas, working with BAIR is a natural partnership,” she says.

“BAIR researchers will have access to great researchers to collaborate with, amazing computing resources, as well as interesting problems with special kinds of data,” Malik says. “Many of the basic ideas of the current AI revolution go back to the 1980s, but computing and data have made today's progress possible.”

The Facebook AI-BAIR partnership announcement follows another achievement for Malik. In April, the IEEE Computer Society announced that he is the recipient of the 2019 Computer Pioneer Award for “significant contributions to early concepts and developments in the electronic computer field, which have clearly advanced the state-of-the-art in computing.” Malik and his research group have developed several well-known concepts and algorithms, including anisotropic diffusion, normalized cuts, high dynamic range imaging, shape contexts, and R-CNN. Malik will receive the award on June 5 at the IEEE Computer Society Board of Governors meeting.

For more details about the BAIR Open Research Commons at UC Berkeley, visit the BAIR website.

To learn more about the IEEE Computer Society's 2019 Computer Pioneer Award, see the official announcement on the IEEE Computer Society website.