Making a Home for all the Parts
Eleven thousand kilometres (or nearly seven thousand miles) lie between Mexico City, where Diana is from, and Sendai, where she currently lives and works as an assistant professor at Tohoku University. That means nine hours of time zone difference, not to speak of the difference in language and culture. I know Diana from organising NAACL 2021, where she was the voice of Latinx in AI at the diversity & inclusion panel and one of the organisers of the Latinx in AI social. Now I meet her on a Saturday that has barely started in Germany and is already reaching its end in Japan, to talk to her about her research trajectory and how she brought Latinx in AI into NLP.
It wasn’t AI that brought her to Sendai though. “I have been doing karate since I was a kid," Diana says. “ I thought it would be awesome to join a dojo in Japan because this is where karate started. That was why I originally learned Japanese.” And neither research nor NLP were on Diana's mind when she finished her undergraduate degree in computer science in Mexico in 2010. “Teachers, policemen and researchers have the worst jobs here”, she says and laughs. But her interest in medicine and life science inspired her to take up a job in data management for a mental health research project at the National Institute of Psychiatry. “What is especially interesting in mental health research, it is not quantitative”, says Diana “You usually ask the patient and they describe how they feel. [...] My job was to take care of the database. Technically it was not my job to look into the data but I am very curious. [...] So I thought, how can we make use of this data? That led me to human computer interaction and that led me to NLP, and I thought: this is what I need to know to use that data. And I was already thinking about doing a masters degree, so let's do that!”
Her inspiration to get into activism followed when she started visiting academic conferences. The Whova app, which was first used at ACL 2019, made it easy to meet outside of conference events and to learn about socials. Queer in AI and WiNLP, both groups that Diana identifies with, were already established, but a group for Latin Americans was still missing. LXAI had a strong presence in ML and CV conferences but not NLP. “We were there but we were not that organised”, says Diana “I saw other people doing it, and I thought we can do it, too.”
Today Latinx in AI takes up its place alongside affinity groups like Queer in AI, Masakhane and Black in AI in NLP conferences, with frequent conference socials, workshops, programs for mentorship and travel support, and a thriving worldwide community. Diana sees lots of commonalities between the groups, and she applies the experience from Queer in AI to her work for Latinx in AI. “We’re aiming for similar things”, she says. “Representation. Seeing senior researchers motivates you. It is important to me as a person to not only do research but to create these spaces and communities. [...] We were already there, but we became visible.”
The fight for diversity continues in day to day interactions in her research lab. “My professor leaves the topic pretty much to me”, she says. While he makes an effort to hire people from all around the world, she doesn’t feel like he engages with diversity beyond that level. “And I see it specifically in my case when we're talking about job opportunities”, says Diana. “I know that he doesn’t mean to give bad advice, but it just shows that he has no idea. He said: ‘you are prepared, it should be easy for you, there are plenty of opportunities for you to find a job.’ […] But particularly in Latin America being a researcher or being a teacher is not good. Having a PhD raises questions like ‘why did you lose your time? Why instead of gaining work experience did you study more?’ There are reasons why we are countries in development. For us this doesn’t make sense. He will need to understand the whole background. He will need to be open to hear my experience. […] There are many researchers who say ‘let’s encourage diversity’ but what are they really doing? How many of them are actually engaging with us?”
Diana doesn’t see herself continuing her career in Japan or becoming a full professor, but rather wants to go into industry. “Staying here I only exist as a researcher”, she says. “The other parts of me are not really relevant. [...] We have different identities. I am a researcher. I’m a queer person. I am a Latin American. I am a woman. All of this is part of who I am. In an ideal world we should be acknowledged not only by the work that we are doing. Humans, we are so diverse. Even if you find someone with the same features that I have, there are things that have a stronger influence on who I am. Who I am is more strongly influenced by me being born in Mexico, more than me being born gay. And there might be a person saying that being queer defines them more than where they were born. That is why we have to talk to people. And that is what I would like to see.” And LatinX in AI is one of the places that makes it happen right now.
You can find more of Diana’s research and writing here
This post was written by Sabine Weber. Sabine is a queer person who just finished their PhD at the University of Edinburgh. They are interested in multilingual NLP, AI ethics, science communication and art. They organized Queer in AI socials and were one of the Social Chairs at NAACL 2021. You can find them on twitter as @multilingual_s