I attended the 2019 INFORMS Annual Meeting in Seattle, WA on Oct 20-23, 2019. INFORMS is the leading international association for Operations Research & Analytics professionals. It’s outside my wheelhouse, but I thought it would be nice to attend since Operations Research does touch on optimization, modeling, and data-driven decision-making, which are topics that interest me.
This was my first time attending an INFORMS conference and I was blown away by the sheer size and scope. There were well over 6,000 attendees and the conference had 100 simultaneous tracks over 4 full days. Compare this to CDC (the flagship annual controls conference), which is about 1/4 of the size.
At INFORMS, I gave two invited talks. The first was in a session organized by Jelena Diakonikolas, who will be my colleague at UW–Madison starting in Spring 2020! My talk was about the similarities and differences between discrete-time and continuous-time dynamical systems and how it relates to algorithm analysis. For example, notions of controllability/observability, stability regions, Lyapunov results, and more have direct analogs in discrete and continuous time. However, notions such as interval stability (e.g. Kharitonov’s theorem) or robust stability (e.g. Aizerman’s conjecture) are notably different in discrete and continuous time! My slides can be found here.
My second talk was in a session organized by Alireza Fallah, a PhD student in Asu Ozdaglar’s group at MIT. This talk was a short version of my talk on unified algorithm analysis and how small semidefinite programs can be used to automate the analysis and design of iterative optimization algorithms. For slides from a similar (but longer) version of this talk, see here.
Overall, I’m happy that I finally got to experience INFORMS. The two sessions I participated in were great (thanks Jelena and Alireza!). I think it’s a particularly nice conference to meet with industry contacts, since there is a substantial industry presence at this conference. That being said, I found it quite overwhelming at times because there were so many talks happening all at once. There is something to be said about smaller venues (see my recent NecSys and Allerton posts!) because the conferences tend to be more cohesive and it’s easier to find your friends when the venue isn’t the size of an airport!