Arjun Subramonian (pronouns: they/them) is a brown queer, agender incoming PhD student at the University of California, Los Angeles. Their research focuses on graph representation learning, fairness, and machine learning (ML) ethics. They're a core organizer of Queer in AI, co-founded QWER Hacks, and teach machine learning and AI ethics at Title I schools in LA. They also love to run, hike, observe and document wildlife, and play the ukulele.
How did Queer in AI begin, and how did you get involved with it?
Queer in AI started at NeurIPS 2017 when a group of artificial intelligence (AI)/machine learning (ML) researchers—frustrated and angry at the lack of community, representation, and inclusion of queer people in AI—created a queer forum on the NeurIPS conference app and met up for coffee. We spent the next year (hundreds and hundreds of emails and many, many meetings!) organizing, creating our mission and platform, and hosting our first workshop at NeurIPS 2018.
I personally got involved with Queer in AI in late 2020. I had previously organized QWER Hacks (Major League Hacking’s first LGBTQIA+ event), for which Luca Soldaini, another Queer in AI organizer, was a speaker. Additionally, I met Willie Agnew, the co-founder of Queer in AI, and Pranav A, also a Queer in AI organizer, at NeurIPS 2020, and all three individuals ultimately convinced me to join Queer in AI. The first Queer in AI events I helped run were the socials at Association for the Advancement of Artificial Intelligence (AAAI) 2021. I had a great time learning the ropes of creating a fun and safe space for LGBTQIA+ individuals to network and socialize. My favorite part about Queer in AI is how supportive all the organizers are of each other. For the AAAI socials, Pranav immediately jumped in and guided me through the process of setting up the registration form, advertising the event, and creating slides to discuss statistics from our demographic survey.
Since AAAI, I've helped establish Queer in AI’s undergraduate initiatives program, in order to engage and support the needs and interests of young LGBTQIA+ folks in AI. The program will help undergraduates get involved in AI research in academia/industry through the support of experienced community members, guide undergraduates to discover their goals and passions, and provide mental health resources and support. We'll achieve this through monthly mentorship events, undergraduate tracks at conferences, and conference buddy programs. We hope that this program, by reaching queer scientists when they’re young, will break down barriers to getting involved with research as an undergraduate and set them up for success when applying to graduate programs.
I'm also organizing the Queer in AI workshop at ICML 2021, which through talks, panels, and group discussions will cover topics like creating safer spaces for trans and non-binary folks in AI, ensuring trans-inclusive publishing, and including non-Western, non-cis gender identities. We also opened a call for submissions related to the intersection of LGBTQIA+ representation and AI or research produced by LGBTQIA+ individuals, and we're accepting submissions in any media (e.g. research papers, poetry, music, etc.) and any language, to maximize inclusivity. Finally, we advise the International Conference on Machine Learning (ICML) diversity and inclusion (D&I) chairs on diversity, equity, and inclusion (DEI) issues related to the conference. For example, we improved the gender and pronouns questions on the conference registration form.
What else has Queer in AI been involved with thus far?
Queer in AI is a decentralized organization, which has helped it tackle a variety of initiatives. Beyond hosting workshops, socials, and panels at nearly every AI/ML/natural language processing (NLP) conference each year, Queer in AI:
- Administers a demographic survey to identify issues within queer communities and shape the future programs of Queer in AI; all questions are optional and anonymous. Through the survey, we ask our community members about inclusion at conferences, mental health, access to research opportunities, and much more! This year’s survey which had 105 respondents found that 80% of respondents reported lack of community, 31% felt their mental health was lowest within the last year, and non-cis respondents gave a rating of 2.6 (on a scale of 1-5) with gender forms. More survey results are listed here.
- Advises conferences on DEI issues. Motivated by the results of our demographic survey, we conducted extensive research into how diverse members of our community would like AI/ML/NLP conferences to improve with respect to DEI and consulted leading scholars in the area. Ultimately, we distilled our findings into a queer-inclusive conference guide which conference organizers can refer to. We are always in the process of improving our content; we have a first draft of our guide on our website that we’re revamping. The guide covers how to: improve registration forms (particularly gender and pronouns collection and privacy), make chat rooms and socials safer, set up a robust reporting system for code of conduct violations, improve speaker and participant diversity, and ensure camera-ready and video guidelines protect trans people from deadnaming.
- Operates the graduate school application review and fee aid program, a joint effort between Queer in AI and our parent organization oSTEM. Our first goal is to end application and test fees as a barrier to queer STEM scholars applying to graduate programs. Because applying to graduate schools typically costs many hundreds of dollars, it can be a major barrier to queer students, who are disproportionately affected by economic hardship; thus, queer students are often forced to apply to a very limited number of schools, take out costly loans, or not apply at all. Our second goal is to guide queer grad applicants in discussing their identities in their personal statements and other application materials; by connecting students with queer researchers in their fields, we provide expert feedback on all parts of grad applications.
- Runs the undergraduate initiatives program I detailed in my response to the previous question
What do you want LGBTQ+ scientists to know about Queer in AI?
Queer in AI is a phenomenal organization to be a part of, regardless of your field of study. Although our events are geared toward supporting LGBTQIA+ researchers in AI, we welcome scientists and humanities scholars from every discipline because our ultimate goal is to foster an inclusive, empowering, interdisciplinary community of queer researchers. We view AI as an interdisciplinary field, and as queer individuals, we're painfully aware of the harms AI systems pose, so we greatly value our members who are experts in neuroscience, disability theory, etc. Furthermore, we are always looking to co-host events with queer communities in other sciences.
Additionally, I highly encourage young and AI-curious LGBTQIA+ scientists to join our organization. You don’t even need to know what AI is; we’re here to help guide you as you explore your passions and interests. Finally, I want to mention that Queer in AI has some of the smartest individuals in the field of AI, and they’re all such kind and caring people; they will put aside whatever they’re working on to help a junior queer scientist in need. This aspect of Queer in AI is truly what makes the organization special to me.
What's next for Queer in AI?
Currently, the primary focus for Queer in AI is just to grow our existing programs while continuing to maintain a large presence at AI/ML/NLP conferences. For example, we're putting on socials, panels, talks, and sessions at NAACL 2021 and hosting the aforementioned workshop at ICML 2021. Simultaneously, we want to get our undergraduate mentorship events up and running with a monthly cadence and bolster the structure of our graduate-application review program.
Has there been anything challenging or surprising that you’ve learned or experienced while doing this work?
I’ve learned quite a few things in the half a year I’ve been an organizer, but I'll focus on two. The first is to lead with self-confidence. While organizing diversity and inclusion initiatives as an undergraduate at UCLA, I would find myself constantly adjusting in response to critical feedback. However, I've learned through Queer in AI that while most critical feedback is valuable, some of it can be misguided, or the result of miscommunication. I now have more confidence in my ideas that are founded in community research, and I frame critical feedback as a response to me not conveying my ideas clearly. The second thing I’ve learned and that has been challenging is how unaware AI research communities and conferences are of issues faced by queer people, and sometimes, how reluctant they are to change their policies, and even registration forms, to better include LGBTQIA+ individuals. There have been a few occasions where I’ve had to carefully opt against burning a bridge in order to continue advocating for our community. This is a major reason we created and are publicizing our queer-inclusive conference guide.
Queer in AI, together with Black in AI and Widening NLP, recently ended its sponsorship relationship with Google, due to the company’s undermining of inclusion and critical research. We collectively wrote an open letter to Google, and this was covered by a variety of news outlets, including Wired.