The Importance of NIPS for Diverse Researchers
Overview
This year, the largest conference on Neural Information Processing Systems (NIPS) will be held in Montreal from December 3rd to December 8th with a variety of sessions, tutorials, workshops, and a competition.
The first significant wave of Artificial Intelligence interest came in the 1950s with Alan Turing, father of modern computing until the 1970s with Norbert Weiner who founded, Cybernetics, the study of control, communication, and organization. Many of the fundamentals of AI came from this timeframe.
It all halted once academic institutions and industry had over-promised and over-delivered on what AI can do at the time. This became known as the First AI Winter. An AI Winter is when interest in AI ‘cools down.’ Therefore, research and funding slow down as well.
However, the interest in AI has sparked once again thanks to the development of AI-powered expert systems in the mid-1980s. It was also during this time where computers became more commonplace in the corporate world and much more powerful. This was known as the second wave of AI. It eventually lasted until the mid-1990s where one of the biggest obstacles was Data Storage, which was expensive at the time.
NIPS began in 1987 as a modest conference on a dark machine learning branch called neural networks. However, it has quickly become one of the world’s most popular conferences for Machine Learning where tickets were sold out in under 12 minutes.
Throughout the past couple of years (2012–2017) there has been a significant increase in interest, primarily due to innovation with ImageNet and Neural Networks.
NIPS 2018 has 8 competition tracks:
- AutoML for Lifelong Machine Learning
- Adversarial Vision Challenge
- The Conversational Intelligence Challenge 2 (ConvAI2)
- Tracking Machine Learning Challenge
- Pommerman
- InclusiveImages: A challenge of distributional skew, side information, and global inclusion
- The AI Driving Olympics
- AI for prosthetics
Diversity
With the implementation of AI into everyday technologies such as web search, social media, and virtual assistants, the uniformity of its developers can lead to outcomes which perpetuate prejudices found in our current society. It is imperative that this technology is free of any biases and is inclusive of culture, gender, ethnicity, and ability of all individuals. Several groups have self-organized to ensure that AI fulfills these goals.
Obstacles
As NIPS has started to diversify these past years, there are many obstacles that these groups faced to make this possible. The burden of diversifying falls on the already marginalized with limited availability as most of the membership consists of full-time researchers. It forces researchers dedicating time to these groups to forego publication deadlines due to their limited availability thus creating a cycle. A cycle where underrepresented groups continue lacking representation as the restrictions of their time and lack of resources hamper the potential quality and availability of their publications. A succession plan and robust organization are needed to ensure researchers are not burdened.
To bring researchers and participants of diverse backgrounds, we need ample time for them to obtain their Visa to enter the country. Another barrier for underrepresented researchers is not everyone can partake in the traditional format of reimbursement from travel grants.
Lastly, as discussed earlier, NIPS registration sells out quickly which decreases the opportunity of diverse researchers to participate in the conference.
WiML
In 2005, four women: Hanna Wallach, Jennifer Wortman Vaughan, Lisa Wainer, and Angela Yu gathered at the NIPS conference, in their excitement to find each other, they made a list of the women that they knew in machine learning. The list only got to 10 women which prompted them to ideate on a gathering for all women in machine learning to connect and share their work. Thus the first workshop focused on Women in Machine Learning was held in 2006, to present and promote their research.
Since 2006, WiML has continued to host their workshop at NIPS presenting on a wide variety of topics such as Bayesian machine learning, Deep Learning, and Reinforcement learning.
Companies that have sponsored WiML in the past include Google, DeepMind, CapitalOne, Microsoft, Facebook, and Amazon.
Black in AI
Another organization breaking barriers for people from Underrepresented backgrounds is Black in AI. Black in AI is focused on increasing the presence of Black people in the field of Artificial Intelligence.
Timnit Gebru, who researches artificial intelligence and is one of the few Black female computer scientists specializing in the field of AI attended NIPS in 2015 for the first time. She counted how many Black people she saw, resulting in 6 for the entire conference of about 8500 people. This made her feel isolated. With the fellow Black researchers she knew, a mailing list was created and a small group planned to have lunch together. This lunch eventually led to talks about a workshop.
The first Black in AI event was co-located with NIPS 2017. Black AI researchers from around the world to shared their work and learned each others’ research. By the numbers:
- The workshop had over 700 friends and Allies
- 450 applicants applied for the travel grant
- 262 abstract submissions
Timnit is trying to counteract the lack of diversity in the field of AI research as part of Microsoft’s honesty, Accountability, Transparency and Ethics in the AI group program
After Black in AI’s first workshop they have had incredible results:
- For those involved, it led to speaking opportunities and startup jobs.
- From a higher education perspective, Cornell’s Ph.D. Computer Science program receive more diverse applicants as a result of sponsorship
- Google Brain started a lab in Ghana
- Internships for researchers
- People that were initially discouraged from studying computer science changed their mind after attending the workshop
- New groups were also created such as Queer in AI and Latinx in AI.
Timnit dedicated 2–3 months full time working on Black in AI to ensure its success.
Latinx in AI
Modeling after the pioneering efforts of WiML and Black in AI, the Latinx in AI Coalition will be hosting the first ever workshop focused on LatinX AI researchers at NIPS in 2018.
Launched in January 2018, 3 Latinx individuals from organizations including the Latino Startup Alliance, Accel.AI and Bay Angels banded together to create a group that would be focused on “Creating Harmony between AI and the Latinx Community.” Seeing how AI has the potential to displace workers of Latinx origin, perpetuate social bias and prejudice and lacks representation of Latinx individuals in the industry, this led us to create the coalition.
Our workshop is a one-day event with invited speakers, oral presentations, and posters. The event brings together faculty, graduate students, research scientists, and engineers for an opportunity to connect and exchange ideas. There will be a panel discussion and a mentoring session to discuss current research trends and career choices in artificial intelligence and machine learning. While all presenters will identify primarily as Latinx, all are invited to attend.
Upon closing of our deadline, we were ecstatic to have over 90 submissions to the first-ever Latinx in AI Research Workshop at the Neural Information Processing Systems Conference.
Submissions have come from top AI Researchers in the US, Central and South America, the Caribbean, Spain, and from Latinx identifying researchers working in other parts of the world!
We are incredibly proud of our progress and hope to continue this growth.
This year at the first ever workshop at NIPS we have selected 20 oral presenters and 40 poster presenters who identify as Latinx from around the world to join us!
See our oral and poster presenters: http://www.latinxinai.org/nips-2018-presenters
Join Us:
If you are interested in sponsoring our initiative please visit our prospectus. We appreciate the support!
We are creating a public directory of LatinX individuals active in artificial intelligence, machine learning, and data science. This directory will be maintained by the LatinX in AI (LXAI) organization.
If you are organizing an event related to artificial intelligence, this list serves as a resource for potential speakers.
Add to our directory — http://bit.ly/LatinXinAI-Directory-Form
Stay Up to Date with LXAI
Subscribe to our monthly newsletter to stay up to date with our community events, research, volunteer opportunities, and job listings shared by our community and allies!
Join our community on:
Facebook — https://www.facebook.com/latinxinai/
Twitter — https://twitter.com/_LXAI
Linkedin —https://www.linkedin.com/company/latinx-in-ai/
Private Membership Forum — http://www.latinxinai.org/membership
