AI is doing tremendous things for our world, like detecting diseases, changing transportation, and helping people connect, but it still has a lot to learn. The more AI learns, the better it will understand the world around us. But equally important is who teaches it: If AI learns from only a small group of people, it might see only a narrow point of view. We believe AI will better serve everyone and advance faster if there is a diverse range of engineers, developers, and researchers teaching it, which is why Facebook AI is working to bring more people from different backgrounds into the field. That process starts in the classroom, where we are working to increase pathways into AI and to build a stronger pipeline of diverse candidates.

Today, we are excited to announce that we are collaborating with U.S. universities that serve significant populations of Black and Latino students so that we can co-teach and fund graduate-level online deep learning courses. We developed and co-taught the first course last spring alongside our longtime partner Georgia Tech, and will expand this program in 2021 to additional institutions and universities with highly diverse populations, including minority-serving institutions.

Facebook’s involvement and interest in increasing underrepresented communities in the artificial intelligence field and investment of resources convinced me to commit to teaching this course. Facebook’s resources have significantly reduced the burden of creating and teaching a high-quality course,” said Zsolt Kira, associate director of the Machine Learning Center at Georgia Tech and an assistant professor in the School of Interactive Computing. “The Facebook lectures provide students with real-world examples and techniques that are needed to deploy and scale algorithms. This is something that students always ask me for, and being able to provide them this information from a large, well-known company that uses these algorithms to process billions of pieces of data per day is invaluable to our students’ education and growth.”

Georgia Tech’s PhD and master’s programs in computer science are among the most successful in the United States at bringing people from underrepresented groups into computing. Moreover, the school’s Online Master of Science in Computer Science program is the largest of its kind in the country, making it an ideal environment for the launch of this initiative.

We are always looking for ways to bring underrepresented groups into the field of computing,” said Charles Isbell, Dean of Computing and John P. Imlay Jr. Chair at Georgia Tech. “This is not only a matter of opportunity for those groups; studies show that more diverse teams solve computing problems better. That is why this partnership between the Georgia Tech Online Master of Science in Computer Science program and Facebook is so important. It will allow us to make a world-class artificial intelligence education available to a wider range of students than ever before.”

Jess Erickson, a Research Program Manager at Facebook, added, “We are thrilled to collaborate with Georgia Tech on the program and to get to know many of their talented and diverse students. We’ve always had a great relationship with Georgia Tech, and with this extended partnership, we will continue to strengthen it and invest in the next generation of artificial intelligence pioneers.”

 

A program designed for real-world applications

Because work in AI advances so rapidly, even world-class graduate programs often teach theories and methods that are no longer used by people at the forefront of the field. It can also be difficult to bridge AI theory and application and show students how to deploy and scale algorithms in the real world. But our inaugural course at Georgia Tech connected the two without a hitch. It combined theoretical and applied AI in a way that resonated with students, earning overwhelmingly positive reviews from those who participated. Students learned techniques and theory with practical applications available each step of the way so that they could easily see how these teachings were applied at scale. In addition, they received mentorship and guest lectures from Facebook’s AI experts and a pathway to interview at Facebook as part of our ongoing efforts to raise awareness of open roles for full-time employment, internships, and residencies.

The semester-long class on deep learning covers fundamentals of neural networks and applications such as computer vision and language understanding. It also enabled students to innovate alongside Facebook AI teams on active projects, including AI Habitat, a simulation platform for embodied AI research, and fastMRI, a collaborative research project to investigate the use of AI to make MRI scans up to 10x faster. Right now, students who are enrolled in the latest course are learning about PyTorch, privacy in AI, and confident machine translation while also engaging in new capstone projects, like Facebook’s Hateful Memes Challenge, an open initiative to advance systems to detect multimodal hate speech. This gives students the opportunity to work on a project that is relevant and extremely important in these unprecedented times.

“By combining the expertise of our faculty and Facebook engineers with state-of-the-art and bleeding-edge content for hands-on experimentation, students are able to engage with these concepts in ways that were previously not possible,” said Irfan Essa, Executive Director of the Machine Learning Center at Georgia Tech.

 

AI belongs to everyone, so let’s teach it together

We’ve seen firsthand that progress comes faster when experts with different perspectives are involved from the start. AI should be researched, built, and deployed by a diverse group of people to ensure that technology is inclusive and reflective of the global community, which is why our partnerships with universities and institutions like Georgia Tech are an important step toward making this a reality. This program is part of Facebook’s long-standing effort to drive diversity and inclusion within the field as we continue to support programs like the Align Master’s in Computer Science, which also seeks to expand the pipeline of people pursuing careers in computer science in the United States, and to invest in organizations such as Black in AI, LatinX in AI, and Women in ML to help contribute to a bigger and broader impact within the community. We’ve also collaborated with organizations like the African Institute for Mathematical Sciences, where Facebook funded and continues to support the African Masters of Machine Intelligence program.

Facebook has adopted a host of programs and initiatives designed to build out the pipeline of underrepresented minorities and women going into tech. We recognize that progress comes not just from sharing research and code, however. A key part of driving fairness in AI is increasing the diversity of the people working on and developing technologies too. That’s why we collaborate on papers with different companies, startups, institutions, and universities. Many of our own researchers are co-employed by Facebook AI and an academic institution, so leading experts don’t have to choose between different settings, kinds of impact, or types of problems to solve.

Likewise, our research scientists who split their time between Facebook, universities, and other academic institutions receive valuable firsthand insights into how academia is pushing boundaries in AI, and they offer learnings from both industry and academia to their students. This distinct perspective, in turn, helps researchers break free of conventional approaches and try bigger, bolder ideas while also teaching their students how to do the same as they begin their own careers in AI. Teaching AI together and giving more people from diverse backgrounds the skills to succeed in the field will serve us well as we continue our fight to build more inclusive and reflective technologies together.

Previous Stanford Materials Scientists Borrow Solar Panel Tech To Create New Ultrahigh-res OLED Display
Next Unifiedpost And Google Collaborate On Document AI To Automate Procurement Data Capture