Motivation and Formation

The operations research (OR) and computer science (CS) communities have recently begun cross-fertilizing their respective strengths in developing foundational methodological and computational tools for artificial intelligence (AI) – enabling decision-making in a broad range of use-inspired problem domains, striving to reach the full latent potential of efforts in individual fields. New methodologies that permit the automated integration of model-building (a focus of OR approaches) and data-driven (a focus of CS approaches) technologies have the potential to combine the best of both worlds, to problems that go beyond the current capabilities of either community. Potential synergies run along multiple fronts, such as combining the computational paradigms and methodological/algorithmic approaches from OR and CS to improve interpretability, trustworthiness, and fairness in decision-making; and combining CS capabilities that can exploit massive data-driven methods with OR capabilities that can capture salient features of complex systems by using a model-based approach to drive effective decision-making.

Professional communities in both fields – namely, INFORMS representing OR researchers and ACM SIGAI, representing CS researchers – that convened in the workshops funded by the Computing Research Association’s Computing Community Consortium, inspired the formation of a dedicated “School” to (i) train the next generation to be dually conversant in both methodological traditions, and (ii) pose challenge problems that are positioned to excite researchers across the CS-OR spectrum, to spur the kinds of advances we believe are possible. We are grateful to the organizers of the CCC Workshops, to INFORMS, and SIGAI for the inspiration.

The inaugural AI-SCORE school is generously sponsored by the National Science Foundation, SIGACT, the University of Maryland Smith School of Business, INFORMS, SIGAI, and AI Journal.

Scope and Topics

Keynote/Overview talks

Keynote lectures by senior researchers will provide overviews of the evolution of foundational OR and AI-related concepts; to introduce bridges that have been built across these fields thus far, and present vision towards future research.

Profs. Michael Fu (University of Maryland), Kevin Leyton-Brown (University of British Columbia), Tuomas Sandholm (Carnegie Mellon University) and David Shmoys (Cornell University) will keynote the inaugural AI-SCORE summer school.

Modules that we focus on in depth

  • Fairness
  • Reinforcement Learning

The The Fairness module will be led by Profs. Siddhartha Banerjee (Cornell University) and Aaron Roth (University of Pennsylvania), and the Reinforcement Learning module by Profs. Vivek Farias (Massachusetts Institute of Technology) and Scott Sanner (University of Toronto). They will jointly create thematic tutorials accompanied by curated datasets and exercises involving hands-on participation by students in cross-disciplinary teams.

Schedule

  • May 27, Day 1: AI-SCORE Opening, Keynotes by overview speakers, Panel Discussion [Michael Fu, Kevin Leyton-Brown, Tuomas Sandholm, David Shmoys]
  • May 28, Day 2: Reinforcement Learning Module Tutorials [Vivek Farias and Scott Sanner]
  • May 29, Day 3: Reinforcement Learning Module (continued)
  • May 30, Day 4: Fairness Module Tutorials [Siddhartha Banerjee and Aaron Roth]
  • May 31, Day 5: Fairness Module (continued)
  • June 1, Day 6: Summary, Feedback and Farewells
Time Talk / Presenter
Day 1 (Monday, May 27, College Park Marriott, Patuxent Room)
8:00-8:30Continental breakfast
8:30-9:00Introductory/Welcome remarks (Profs. Lavanya Marla & Ferdinando Fioretto, Chairs)
9:00-10:30OR Perspectives to OR/AI Integration: Optimization lens (Prof. David Shmoys)
10:30-10:45Coffee Break
10:45-12:15OR Perspectives to OR/AI Integration: Stochastics lens (Prof. Michael Fu)
12:15-1:30Lunch
12:15-1:30Self-introduction of participants
1:30-3:30CS Perspectives to OR/AI Integration (Prof. Kevin Leyton-Brown)
3:30-3:45Coffee Break
3:45-5:45CS Perspectives to OR/AI Integration (Prof. Tuomas Sandholm)
6:00-7:30Dinner
7:30-8:45Panel discussion (Profs. Ferdinando Fioretto, Sven Koenig, Kevin Leyton-Brown, Lavanya Marla, Tuomas Sandholm, David Shmoys)
Day 2 (Tuesday, May 28, Van Munching Hall room 1330, Smith School of Business)
8:00-9:00Continental breakfast
8:30-9:00Welcome (Senior Associate Dean Wedad Elmaghraby, Prof. Balaji Padmanabhan, Director of AI Center; University of Maryland Robert H. Smith School of Business)
9:00-10:30Reinforcement Learning (Prof. Scott Sanner)
10:30-10:45Break
10:45-12:15Reinforcement Learning (continued)
12:30-2:00Lunch
1:30-2:00League of Robot Runners competition (Prof. Sven Koenig)
2:00-3:30Reinforcement Learning (Prof. Scott Sanner)
3:30-3:45Break
3:45-5:15Reinforcement Learning (continued)
5:30-7:30Dinner
Day 3 (Wednesday, May 29, Van Munching Hall room 1330, Smith School of Business)
8:00-9:00Continental breakfast
8:30-10:30Reinforcement Learning (Prof. Vivek Farias)
10:30-11:00Break
11:00-12:30Reinforcement Learning (continued)
12:30-2:00Lunch
2:00-3:30Reinforcement Learning (Prof. Scott Sanner)
3:30-3:45Break
3:45-5:15Hands-on exercises
5:30-7:30Dinner
Day 4 (Thursday, May 30, Van Munching Hall room 1330, Smith School of Business)
8:00-9:00Continental breakfast
9:00-10:30Fairness in Resource Allocation: Axioms, Optimization, and Markets (Sid Banerjee)
10:30-10:45Break
10:45-12:15Introduction to Fairness in Machine Learning (Aaron Roth)
12:30-2:00Lunch
2:00-3:30Fairness Tradeoffs in Online Allocation Settings (Sid Banerjee)
3:30-3:45Coffee Break
3:45-5:15Multigroup Fairness in Uncertainty Quantification (Aaron Roth)
5:15-5:30Break
5:30-7:30Dinner
Day 5 (Friday, May 31, Van Munching Hall room 1330, Smith School of Business)
8:00-9:00Continental breakfast
9:00-10:30Multicalibration and Downstream Decision Making (Aaron Roth)
10:30-10:45Coffee Break
10:45-12:15Fairness topics (Sid Banerjee)
12:30-2:00Lunch
2:00-3:30Discussion Sessions (Fairness Case-Studies): price discrimination, organ donation, bandits
3:30-3:45Coffee Break
3:45-5:15Breakout sessions and wrap-up
5:30-7:30Dinner
Day 6 (Saturday, June 1, College Park Marriott, Room 1308)
8:30-9:00Continental breakfast
9:00-11:00Student's research discussion
11:00End of the school (lunch on your own)

Eligibility

PhD students who are entering their 2nd year of PhD in Fall 2024, or who are just completing their 2nd year of PhD in Spring 2024 are eligible. Students from underrepresented communities are especially encouraged to apply. Only 15 students each from the fields of OR and CS/AI will be selected due to space constraints.

How to Apply

Please apply at the Google form here. The form will request the student’s CV, a one-page (maximum) statement describing the problem area at the intersection of OR and CS/AI being explored by the student as a PhD topic; and a short statement from the PhD advisor confirming the student's research area in OR/AI andthe anticipated benefit (maximum one page).

Application deadline: May 4, 2024

Advisory Committee

  • Michael Fu, University of Maryland College Park
  • Sven Koenig, University of Southern California
  • David Shmoys, Cornell University

Organizers