Having a diverse group of individuals shaping the future of AI is crucial. People with diverse experiences and backgrounds are able to bring unique ideas and questions and approach the technology in new ways, which ultimately improves the quality of our work and ensures that it serves everyone effectively. Furthermore, it is our collective responsibility to make sure that the people shaping the future of AI represent the full breadth of society, so that everyone has a say in the direction of the field.
With these goals in mind, Facebook hosted a Women in AI Summit in November to strengthen our internal community and discuss how to foster greater diversity and inclusiveness. The event brought together Facebook AI’s women engineers, research scientists, administrators, program managers, and emerging scholars. The summit featured a keynote from University of California, Berkeley, Computer Science Professor Dawn Song, which followed a welcome presentation from Facebook VP of AI Jerome Pesenti and Facebook AI Research (FAIR) Managing Director Joelle Pineau.
Pesenti talked about the current state of diversity in AI at Facebook and what they hope to achieve. “I’m concerned that the numbers aren’t where they should be, but I’m hopeful for the future — that we have the tools, resources, and people to do better,” said Pesenti in his welcome presentation. “To get there, I’d like to ask for everyone’s help in fostering inclusion within AI at Facebook.”
Next year, the Women in AI Summit will expand to include external attendees throughout industry and academia. We will also extend our focus to include underrepresented groups representing Latinx in AI, Black in AI, LGBTQ in AI, and others. “We’re looking forward to seeing how the community grows and evolves,” says Jess Erickson, AI Diversity and Inclusion Program Manager. This summit was a necessary step in the right direction, and it’s important to recognize the hard work of our steering committee, task force, and summit participants.”
To highlight some of these efforts, we sat down with four leaders of diversity initiatives within Facebook AI, Emily Dinan, Danielle Rothermel, Nikhila Ravi, and Michelle Cheung, to learn more about what motivates them, why having community matters, and the extra time they’ve dedicated to building the Women in AI community at Facebook.
Emily Dinan is a Research Engineer in the Facebook New York office. Her current work focuses on dialogue research, or training machines to have conversations. With others in the FAIR NYC office, Dinan has helped lead the New York chapter of the Women in AI group, and she’s organized meetups with research talks from various women across Facebook AI. In addition, she also hosts casual monthly lunches in the office for women in AI. “The lunches have been the most rewarding for me,” she says. “Last year, I had a couple of interns thank me for organizing them, because they never had the opportunity to talk to other women during their summer internships. These lunches were their first chance to do so.”
Hiring diverse talent is crucial, but retaining that talent is just as important. “Having a community like Women in AI contributes to your sense of belonging, and feeling like you belong then contributes to job satisfaction and retention,” Dinan says. “In industries that are male-dominated, women can spend entire days where they’re the only woman in a meeting, and sometimes their team, managers, and reports are all men. It can be isolating. This is why it’s so important to have a community to talk and share experiences.”
Danielle Rothermel is a Research Engineer who is currently doing research on spatial understanding in Minecraft. In the past, she has worked on StarCraft-related research in FAIR. Rothermel is an active member of the Women in Engineering community and has participated in a mentorship program since joining Facebook. She has also attended recruiting events targeted toward both women and black AI researchers and engineers, contributed to an internal Women in AI spotlight (organized by Nikhila Ravi), and participated in this year’s Women in AI Summit panel with several other women researchers. The panel discussed how researchers and others can grow their professional networks, mobilize other women to pursue careers in AI, build a greater sense of belonging, debug challenging moments at work, grow their leadership skills, and more.
“Diversity is something that I very much care about,” Rothermel says, “but there are so few black engineers and so few women engineers in the industry in general that, sometimes, if I don’t participate in a panel or initiative, there might not be anyone to represent my group. So while I definitely do want to participate, there is a lot of pressure placed on ‘doubly underrepresented’ people in particular.”
Rothermel emphasizes the importance of showing up for your community. “It’s encouraging seeing yourself in the place you want to be,” she says, “but what about once you get there? It helps to see a community of other people who have also made their way to a similar place as you — and seeing that you’re not the only one.”
Nikhila Ravi is a Research Engineer working on computer vision. Since joining FAIR, she has worked on a project that used deep learning to take 2D key points of objects and people and lift them to 3D; it was presented at ICCV 2019 in Seoul. Ravi is on Women in AI steering committee in Menlo Park, and helps organize opportunities for women in the field to come together and connect. She also runs a monthly internal spotlight series designed to highlight experiences and share advice from women in different AI roles. According to Ravi, sharing our experiences facilitates connections, which helps develop and strengthen bonds within a community.
Ravi is motivated by the experiences of her mother, who was a systems engineer at Bell Labs in New Jersey in the 1980s. “She was working on cutting-edge research at the dawn of the internet,” Ravi says, “but she only stayed at Bell Labs for one year before moving back to India. She had a lot of potential, and if she’d had strong allies to provide support and encouragement to develop her career, she could have made a real impact.
“Since FAIR is one of the leading industry AI research labs, I want to do my part in ensuring that all women feel included and valued and don’t miss out on becoming leaders in the AI revolution,” she says.
Michelle Cheung is a Product Designer on the AI design team, where she works closely with the applied AI research team supporting computer vision and video understanding. “I take their technology and try to inject a more human perspective to it, keeping in mind what the needs of the user are,” she explains. Cheung, along with Ravi and others, was one of the first members of the Women in AI steering committee in Menlo Park. After joining forces with the Women in AI group in New York, the global team (led by Erickson) plans to grow local chapters in the other AI offices by providing guidebooks.
“Hiring more women and women of color is a top priority, and having this group will help provide support for potential new hires by showing what it’s like to work at Facebook AI from a woman’s perspective,” Cheung says. A self-described introvert, Cheung knows the importance of facilitating connections. “I want to encourage people to know each other, to learn from each other’s strengths and weaknesses. When you encounter a problem, someone who’s been in the same shoes as you could provide valuable insight into your situation.”
Those looking to connect with our growing Women in AI community can find us at any upcoming Women in CV or Women in ML (WiML) workshop. Our Women in AI researchers frequently attend conferences like ICML, CVPR, ICCV, EMNLP, and more. Come see us at the upcoming WiML workshop at NeurIPS on December 9, and be sure to follow us on Twitter @facebookai for updates.