This article is published in the November 2014 issue.

Broadening Participation in AI

At AAAI-13, several attendees remarked that the number of women and underrepresented minorities in attendance seemed even lower than in previous years, and they started talking about what could be done. So for 2014, with the encouragement of AAAI and financial support from the CRA-W and Coalition to Diversify Computing (CDC) through the Discipline Specific Workshops program, we – Maria Gini, Adele Howe, Monica Anderson, and Andrea Danyluk – organized a set of activities aimed at increasing the number of women and members of other underrepresented groups in AI by encouraging and mentoring students and post-docs.

The activities were held at AAAI-14 in Quebec City, Quebec, Canada and spanned the entire conference: July 27-31, 2014. The sessions were designed to provide career mentoring, build community, and give participants the opportunity to network with peers and established researchers. The program also provided registration and travel support to many participants who needed it.

Twenty-one students and post-docs participated, including 16 women and 5 men. Nine participants self-identified as African American or Hispanic.Beyond the organizing committee, a dozen AI researchers volunteered as panelists and mentors.

Building a Community

Building on a similar model for a workshop held at the Autonomous Agents and Multiagent Systems (AAMAS) conference in 2013, the AAAI-14 Discipline Specific Workshop began with an afternoon session the day before the start of the main conference program. The first part of this session comprised introductions and a career panel. Panelists discussed and answered questions on a wide range of topics, including the process of picking a research area and advisor, tips for networking, and research and teaching careers post Ph.D. The remainder of the session was devoted to small-group mentoring. Participants met in groups of two or three with established researchers in related subareas of AI to discuss research and get advice.

Beyond being an informational and mentoring session, this segment of the workshop was designed to build community. Participants had plenty of time to meet each other, and the mentors encouraged them to take advantage of and seek out opportunities to learn and network at the conference. In a post-workshop survey, participants said they felt welcomed and supported by the AI community as a whole and enjoyed the opportunity to build their own communities and networks within it.

On the first formal day of the conference, the participants attended a lunch (sponsored, in large part, by the AI Journal) to which all women at AAAI had been invited. This significantly broadened their exposure to members of the AI community. Participants and mentors felt that this helped reinforce the sense of cohesion and community and enjoyed lively discussions during the meal.

On the final day of the conference, the participants gathered for a breakfast, sharing new insights they had gained at the conference and providing feedback to the organizers.AAAI DSW Photo

Building a Career

Many factors contribute to the success of a research student or post doc. (1) The student needs to understand previous research as well as current research to contextualize her/his work. (2) S/he needs to develop an independent research project. This requires a creative mind; it also requires the opportunity to get feedback and suggestions not just from the student’s advisor but from others in the field. (3) The student needs to build a network. (4) The student needs to build confidence in her/his abilities to ask questions, to explain her/his research ideas and problems and to contribute to the field.

The AAAI-14 Discipline Specific Workshop was designed to provide elements in all of these areas. As one student said in a post-workshop survey, s/he would not have attended AAAI without the support of the workshop. Thus the impact the workshop had on enabling students to attend the conference was critical, giving them the opportunity to learn about cutting-edge research, talk to the researchers doing the work, and get feedback on their own research from the workshop mentors as well as from other conference attendees. As another participant said after the conference, “[the] mentoring session was by far the highlight, and I’d love to repeat that.” Finally, the workshop gave students the opportunity to network with the wider AI community. As one participant noted, “Contacts that I established may be useful in helping me find a position in academia.”

Lessons Learned

Overall, participants who answered the survey question “Would you recommend this workshop to others?” unanimously said they would, but they also had suggestions to make the workshop even better. As organizers of the workshop, we hope that our own experience, together with the participants’ feedback, can serve to provide others with a starting point for running similar “broadening participation” workshops.

If co-located with a conference, hold the first workshop session before the conference’s official start. For some participants, this will be their first time at a research conference. They might be intimidated or simply not know anyone in the community. The workshop, then, is a perfect setting for participants to begin networking and also to get advice on how to get the most out of the conference.

If possible, arrange small group mentoring sessions. Some students who are new to a research area may be intimidated to discuss their work. Small group sessions give students the support of other students in the group, while being intimate enough to encourage discussion. Match the research interests of the mentors directly to the research interests of the students for maximum mutual benefit.

If co-located with a conference, spread the workshop sessions over the duration of the conference. Periodic “check-in” with participants allows them to re-connect with each other and to address questions they may be accumulating over the course of the conference.

Remember that there are distinct stages in the lives of research students and post-docs. The progression from new student through Ph.D. through post-doc is a long one. Advice that is valuable to students early in their careers isn’t the same as the advice that’s valuable later on. Aim to address broad topics in large-group settings and more narrow topics in small-group discussions.

Support for Discipline Specific Workshops

To learn more about the CRA-W/CDC Discipline Specific Workshop program, including how to apply for funding, go to:

About the authors:

Andrea Danyluk, Maria Gini, Adele Howe, and Monica Anderson are professors of Computer Science at Williams College, the University of Minnesota, Colorado State University, and the University of Alabama, respectively. Their research interests are in Artificial Intelligence. Andrea Danyluk and Maria Gini are members of CRA-W. Monica Anderson is a member of CDC.