Since 2008, the Women in Machine Learning (WiML) Workshop has been held in conjunction with the Conference on Neural Information Processing System (NeurIPS), the foremost artificial-intelligence conference. The WiML workshop is the flagship event of the WiML organization, whose mission is to increase awareness and appreciation of the achievements of women in machine learning (ML).
According to Hanna Wallach, one of WiML’s cofounders, the organization was born in 2005 when she and three other women found themselves sharing a room at NeurIPS and “talked about how exciting it was that there were four female students at [the conference] that year.” This year, the WiML workshop’s mentorship round table alone has more than 50 senior ML scientists volunteering as mentors, in addition to the workshop’s usual complement of keynote addresses, research talks, and panel discussions.
The general chair of this year’s WiML is Sergül Aydöre, a senior applied scientist with Amazon Web Services, who took a break from workshop planning to answer three questions for Amazon Science.
- Q.
What are the WiML workshop’s goals?
A.The workshop was created to increase the visibility of women in the field and to give us an opportunity to help each other with the challenges that women face.
The workshop celebrates women’s achievements. We make it more obvious that there are many successful women and nonbinary people in this field. We invite them to give technical presentations about their fields, so that women in their early careers can learn from them and get inspired. And we ask them to share their experiences. Because it is different for women, for sure.
We also want to give women a chance to network. In ML, as in any career, networking is quite important. When you have networks, then you can go and give a talk about your paper or your work, and when you're connected, it makes it much easier to get a job. Or internships — when I'm there, I can say that we are hiring interns for next summer and that people should reach out to us if they are interested.
- Q.
How does the workshop promote those goals?
A.In the technical talks, the content is not too much different than the content for the general conference. We try to make sure that we don't have speakers who are already speaking in the main conference. Anybody who comes to this conference can find our technical content relevant; all genders are welcome in these technical talks.
When you have networks, then you can go and give a talk about your paper or your work, and when you're connected, it makes it much easier to get a job.
Sergül AydöreWe also have mentorship sessions. Some of them are on specific technical areas like natural-language processing, but we also provide mentorship on topics like choosing between industry versus academia, how to have a more balanced life, or life with kids.
We have a specific time for mentorship roundtables, with tables for each topic. We ask mentors for that topic to sit at the table, and then attendees can just sit at any table that they are interested in.
Another thing we try to do is to ensure geographic diversity. This year, for example, we tried really hard to make sure that we have geographic diversity for the speakers that we invited, and we also support travel funding for students all around the world. We also provided visa letters for those who are coming to the US from different countries, in order to help them with their visa applications. We try to do everything we can to make sure that they can actually come and attend and learn from other women. And we thank our sponsors — including Amazon Science — for the financial support they provided.
- Q.
This year is the 15th year that WiML has been held at NeurIPS, after two prior instances at other conferences. Do you think it’s been a success?
A.I absolutely think so. This is one of the most respected workshops at NeurIPS, because it has been going for a long time and the quality is pretty high. Of course, the pandemic temporarily reduced attendance, but up to the pandemic, the numbers had been steadily increasing since the first time the workshop was offered.
What really makes me happy is that, for example, we invited senior researchers to give invited talks, and most of them mentioned that their first presentations in ML were actually at the WiML workshop. When they were students, they attended WiML as presenters, and now they are inspiring other students. I think that is really nice to see.
That’s true for me, too. When I first attended in 2017, I didn't have any papers published in a main AI conference until then. I had a poster presentation in WiML, and I was so excited to present it at such a big conference. That was my first time really interacting with AI researchers about my work. I got very useful feedback on my work during the workshop, and then that work was published in a main conference later.
If you just look at the successful women researchers in AI who are presenting this year, where they are at their careers, and where they were when they first interacted with AI through WiML, I think that shows the success of the workshop.
I am also very impressed by the quality of the abstracts we received. We selected some of these works as oral presentations, and the rest will be presented as posters. Of course, we could not do it without the help of our area chairs, who reviewed the submitted abstracts carefully. These area chairs, who are also women or nonbinary, were recruited based on their expertise in AI. Additionally, we had women volunteers to help us out during the event. Given that all the efforts in every aspect of our workshop are voluntary, it is so nice to see this collective effort of women helping women.
I encourage all NeurIPS attendees to visit the WiML workshop and celebrate the achievements by women in ML.
