WNCG Faculty Evdokia Nikolova Co-Organizes Program on Real-Time Decision Making

Wednesday, March 07, 2018

WNCG faculty member Evdokia Nikolova is co-organizer for a semester-long program on Real-Time Decision Making at UC Berkeley. The program is hosted by the Simons Institute for the Theory of Computing, a collaborative research institute which is a leader in theoretical computer science.

Scientists investigating topics from astrophysics and earth science to urban mobility and energy work with vast amounts of data. The issues they are trying to address have practical outcomes in our day-to-day lives: creation of better earthquake warning methods, observation of astronomical phenomena, improvement of transport infrastructure, and optimization of energy grids, to name a few. One challenge across the board, however, is finding a way to use the data that have been collected in a methodical way that can accurately predict an outcome to a given problem. That’s where theoretical scientists like Professor Nikolova come in.

If the numerical data obtained from real-world scenarios could be translated into a mathematical algorithm, it could mean great advancements for these fields. However, representing real-world situations in a mathematical way isn’t always straightforward.

“In the past five years, I’ve made a deliberate effort to connect to colleagues and researchers from applied disciplines outside of theoretical computer science […] in search [of] interesting algorithmic questions,” Nikolova said, speaking in the program’s kick-off workshop, “Real-Time Decision Making Boot Camp.” Yet in her conversations with these colleagues, she found there was something of a disconnect between the applied scientists and the theoretical scientists like herself.

Scholars in different disciplines often use terminology specific to their field, and they also have distinct ways of visually representing their data. The lack of a mutually understandable way to talk about the issues at the heart of key research questions inevitably limits the chance at successful collaboration.

“I’d think, ‘Well, I don’t really understand what these problems are…in an algorithmic way,’” Nikolova recalled.

The main goal of the Simons Institute’s Real-Time Decision Making program is to bridge that gap—to get participants in all of these disparate fields on the same page from the beginning. Once common ground is established, then a more productive discussion on existing problems and possible modes of solution can take place. The program was designed with the Simons Institute’s larger goal in mind: to “promote fundamental research on the foundations of computer science, as well as to expand the horizons of the field by exploring other scientific disciplines through a computational lens.”

As a co-organizer, Nikolova joins Professors Josh Bloom (UC Berkeley), Steven Low (CalTech), Balaji Prabhakar (Stanford), and program chair Richard Karp (UC Berkeley) in planning the program, which is one of two being hosted at the Simons Institute’s Calvin Lab this semester. Highlights of the program include the aforementioned Boot Camp followed by three weeklong workshops. The Boot Camp session, which took place January 22 through January 26, provided a quick overview of basic concepts in the fields represented by the long-term participants. Nikolova presented a “Brief Introduction to Algorithms, Game Theory and Risk-Averse Decision Making.”

Aside from the workshops, the long-term participants of the program are fully immersed in their research. Since cross-field collaboration is the highlight of the program, they have unstructured days during which they are encouraged to set up meetings with other participants. They also have “tea time” every day—a more casual chance to make connections, catch up on each other’s studies, and foster collaboration naturally.

The program includes almost 50 long-term participants as well as 19 Visiting Graduate Students and Postdocs. Eight Research Fellows round out the list. WNCG student Ali Khodabakhsh, a Graduate Research Assistant in Nikolova’s research group at the University of Texas, is among the attendees, as is WNCG Postdoctoral Fellow Dr. Emmanouil Pountourakis, who won a Simons Fellowship to attend the program.

Real-Time Decision Making” is running at UC Berkeley’s Calvin Lab from January 9 through May 11. Registration for short-term participation in the workshops is free and open to the public. Upcoming workshops in the program include “Applications in the Natural Sciences and Physical Systems,” “Societal Networks,” and “Mathematical and Computational Challenges in Real-Time Decision Making.” For those who can’t make it in person, each workshop will be recorded for live-stream, and video will also be accessible after the event.