March 8, 2018

Why diversity matters in AI research

By: Joelle Pineau

Dr. Joelle Pineau is head of the Facebook AI Research (FAIR) lab in Montreal and co-director of the Reasoning and Learning Lab at McGill University’s School of Computer Science. On International Women’s Day,Pineau discusses the importance of diversity in the field of AI research and the progress being made to bring more young women into the field.

AI is already a regular presence in our daily lives, powering everything from the music playlists we stream to the product recommendations that show up when we shop. As the technology continues to evolve, it’s important that we have a diverse set of people designing and creating it. This includes people with diverse experiences and personal characteristics to ask different questions, see the technology differently, and design differently.

I regularly give talks, or speak on panels, aimed at people outside of our field. These general outreach efforts are not so much intended to bring more people into the field, but rather to open up a discussion with the general public about the development and impact of AI research. It’s an opportunity for us to demystify what we do, and it’s important to me to understand their experiences, challenges, and ideas for changing the world. We need all of these perspectives to make sure the technology meets the needs of everyone.

Dr. Joelle Pineau explains AI to a non-expert with a help of a dog.

AI research is so much fun, and there are so many interesting, exciting opportunities. I want those opportunities to be open to everyone, regardless of gender, background, ethnicity, or any other personal characteristic. Diversity, of course, is core to Facebook’s business. It informs the products we build, the features we improve, and the people we hire. Still, within the wider field of AI research, while we’ve made some progress, women and under-represented minorities are not yet well represented, and women in particular leave the field at a higher rate than men at every career stage. We must all do a better job of communicating how diverse and fascinating working in AI can be to encourage a wider spectrum of people to enter the field. And we need to nurture the careers of young scientists, and provide a work environment that is interesting, safe, and rewarding, for all members of the community.

Encouraging young women to study computer science

The first goal is making it easier for young women to discover the field of computer science in college. At McGill University we have started to tackle the issue of gender equality in the field. Our first goal was to make it easier for young women to discover the field of computer science in college, and we have significantly increased enrollment of women in our undergraduate CS programs by having several multidisciplinary programs, such as CS+Math, CS+Bio, CS+Arts, and CS+CognitiveScience.

The other way we have increased enrollment is by making it easier for students to switch into a CS program at any point in their undergraduate career. A few years ago, we ran a survey that revealed that many of the male students in our graduating class knew from day one that they wanted to be CS majors. In contrast, most of our female graduates decided to major in CS in the second or third year of undergrad. They started by taking one course, which they liked, then took a few more and realized they were good at it. By having a program that allowed several entry points, we effectively open the doors to a more diverse student population.

Bringing women together inspires progress

Building diversity in AI is crucial, but it’s just as important to continue to nurture young scientists. One way to do this is to recognize the work of women researchers, and to communicate the potential that their work has to influence the field. I often participate in events like the annual Women in Machine Learning (WiML) event, which started over 10 years ago, and I started the Women in Robotics workshop at McGill a few years back. At Facebook I mentor researchers and seek to create opportunities to bring together young scientists with industry researchers for networking and mentorship building opportunities, such as a Women in AI event I am hosting later this month in Montreal. I find these events very refreshing, and love seeing brilliant, inspiring women explain their research. It is always wonderful to meet younger researchers, PhD students and postdocs who are beginning their careers. And it’s important to me to understand their experiences, challenges, and ideas for changing the world, and to share some of my experiences with them.

As a researcher, I have collaborated with research teams in neuroscience, anthropology, medicine, social media, law, solving important problems at the intersection of these disciplines and AI. There is an opportunity to dive into so many interesting problems that can have a real impact in people’s lives. AI is a field that requires creativity, imagination, good problem-solving skills, and strong communication skills. A love of math & science helps too; computer programming can be learned much more easily than most of these other skills.

The field of AI is very competitive, and there is a belief that “merit” is the only criteria for success. But the reality is more complex than that. Biases are hard to recognize or confront, and it is often much easier to appreciate creativity, intelligence, and hard work in someone who looks like us, behaves like us, and comes from a similar background. There is much that we can do to be more receptive to talent, impact and contributions that come from people that are different than us. We need all of these perspectives to make sure that the technology meets the needs of everyone.

My advice for women who want to get into AI research? Be passionate and be curious about your passions. Seek opportunities to collaborate with others. Start with one course and see where it takes you.