INFORMS AI Initiative and the AI-OR Workshop

Ramayya Krishnan launched the INFORMS Artificial Intelligence (AI) initiative in 2019 during his term as president of INFORMS, with the appointment of an ad hoc AI Strategy Advisory Committee (AISAC). Several key outcomes resulted from that initiative, including the publication of an INFORMS whitepaper on AI [1], a plenary presentation on INFORMS AI Strategy [2] at the 2019 INFORMS Annual Meeting, creation of a vision document submitted to the CCC (Computing Community Consortium) [3], the launch of the new INFORMS Journal on Data Science (IJDS) [4] and the release of the INFORMS Analytics Collection (formerly Editor’s Cut) titled, “Advances in Integrating AI and O.R.” [5].

At the beginning of 2021, INFORMS President Steve Graves requested that I establish an ad hoc AI Strategy Workgroup to continue the efforts that began in 2019. The group includes Jake Abernethy, Sudip Bhattacharjee, Sanmay Das, John Dickerson, Jay Shan, Phebe Vayanos, Pascal Van Hentenryck and Elena Gerstmann, with Krishnan continuing his valuable guidance on this initiative.

Briefly, INFORMS’ goals with the AI initiative include the following:

• Advance and promote O.R. and analytics within the AI community
  • Elevate work of INFORMS members within the AI space
  • Explore efforts in education, publications, workshops at the intersection of O.R. and AI to increase the awareness of AI within O.R. and vice versa
  • Collaborate with other communities in AI (ACM, SIAM, ORS, EURO, CCC)
  • Support outreach to industries and enterprises: Can practitioners learn about the intersection of O.R. and AI through INFORMS? Can companies turn to INFORMS and O.R. professionals when hiring AI talent?
• Advocate for O.R. and analytics among policymakers and other constituencies
  • Increase funding opportunities in NSF and other sources (DoD, OSTP, DOE, DOT, etc.) related to AI initiatives
  • Position INFORMS members as a resource on AI initiatives
  • Educate policymakers around the role of O.R./analytics within AI

Fortuitously, INFORMS’ initiative coincides with the renewed emphasis on AI at the federal government level and its National AI Initiative. The ad hoc group is working closely with Jeff Cohen, Beril Toktay and the Signal Group [6] to ensure that we stay in tune with the national directives and opportunities.

Some of the activities undertaken by the ad hoc group in 2021 include:

1. Correspondence/meetings with policymakers and leaders in OSTP and NSF
2. Alerting CAI, DM, QSR communities and individual INFORMS members about AI-related RFIs and nominations from the federal government
3. Promoting examples of using O.R. and AI to solve critical, societal problems

AI-OR Workshop

The most significant event under this initiative in 2021 was the first workshop (in a three-event series) on AI-OR, jointly organized by the CCC, ACM SIGAI and INFORMS [7]. This workshop took place virtually on Sept. 23-24 and provided a space for AI and O.R. communities to discuss a joint strategic vision for a strong and sustained collaboration between the two fields. The goal of the workshop was to bring together these communities to review the state of O.R. and AI along multiple axes, articulate research and education opportunities as well as grand challenges of mutual interest, and explore topics for summer schools and research programs to foster AI-OR collaborations.

This workshop included eminent speakers from both the O.R. and AI communities and covered methods and applications. Presentations were followed by Q&A and breakout group discussions among the participants. The workshop drew more than 60 participants with equal representation from the AI and O.R. communities, with a mix of academia, industry and government representatives. A report out [8] from this first workshop was completed just in time to be shared at the 2021 INFORMS Annual Meeting in Anaheim, Calif., and both communities shared it with their members via social media posts, etc. The report summarizes presentations, breakout discussions and key takeaways from each session at the workshop. Video recordings of the presentations are linked in the report. The second and third workshops are expected to take place in the first half of 2022. 

Briefly, the first workshop participants:

• Dispelled common misconceptions across the two communities and discovered common objectives, albeit a few different assumptions leading to different tools and methods
• Discussed multiple opportunities for collaboration at the data, methods and policy layer
• Proposed potential future directions, such as the availability of datasets, new competitions geared toward the AI and O.R. communities, an AI/OR summer school for Ph.D. students, possible AI-OR Ph.D. programs, a possible ACM/INFORMS AI/OR conference or journal, and curating examples and case studies that combine AI and O.R. to achieve better results

In summary, the consensus is that multidisciplinary collaboration across the two communities (and other relevant communities) is essential for continued innovation in O.R. and AI. We plan to continue these discussions in future workshops and propose some joint actions that can be undertaken in support of this collaboration.

References and Notes

1. https://cdn2.hubspot.net/hubfs/3449182/INFORMS_AI_Whitepaper.pdf
2. https://youtu.be/5beyzv3Mjfg
3. https://cra.org/ccc/
4. https://www.informs.org/About-INFORMS/News-Room/Press-Releases/INFORMS-Launches-New-Journal-on-Data-Science-to-Fill-Critical-Gap-for-Data-Science-Researchers-Professionals-and-Decision-makers
5. https://pubsonline.informs.org/editorscut/ai-ml
6. https://signaldc.com/our-people/
7. Workshop organizers include: Sanmay Das, John Dickerson, Pascal Van Hentenryck, Sven Koenig, Ramayya Krishnan, Radhika Kulkarni and Phebe Vayanos, with support from Ann Schwartz, Daniel Lopresti and Maddy Hunter from CCC, and Elena Gerstmann from INFORMS.
8. https://cra.org/ccc/wp-content/uploads/sites/2/2021/10/Filler-AIOR-document.pdf