Perhaps Paul Roebber, UWM Professor of Atmospheric Sciences, is best known for founding the Innovative Weather Service (see related story above). That program, however, represents just one facet of his expertise. Roebber also has background in machine learning, agent-based modeling, and artificial intelligence. In this interview, he shares how his knowledge and skills can be applied to real-world situations. Be sure to read all the way to the final fun fact. Because, well, it’s fun.
Paul Roebber,Distinguished Professor,
School of Freshwater Sciences
What kinds of projects are you working on now, and what led to them?
I recently finished a project with the National Oceanic and Atmospheric Administration (NOAA), which
oversees all the centers, labs, and individual offices of the National Weather Service. They asked me to
perform a review of machine learning activity across all these entities and identify obstacles to future
progress.
That work led to some follow-up projects, which could be commercially interesting. One is developing a
flash-flood agent-based model that will assist in emergency management. Another project applies this
type of modeling to fire weather disasters – forecasting conditions that are conducive to ignition, spread,
and behavior of wildfires. I am especially concerned with wildfires, because while we think of those
events as being confined to the western U.S., that is no longer the case. I worry about the Wisconsin
Northwoods, for example.
What are some examples of ways you can help companies?
I think there is some really untapped territory for insurance companies. Losses can arise from all manner
of severe weather events, including wildfire. Modeling can help understand the risk going forward and
help mitigate such losses. There are already catastrophe modeling companies out there, but they aren’t
capturing the full scope of variables that I believe to be impactful. For example, we need to understand
the changes in frequency of all kinds of severe weather, and where places that did not experience those
kinds of conditions will now have to deal with them. Wisconsin is not immune.
Tell us one fun fact about you personally or professionally.
I have applied my modeling expertise to the world of professional sports and published articles to
answer burning questions such as:
• Does the principle of investment diversification apply to the starting pitching staffs of major
league baseball teams?
The answer is, surprisingly, no. Bet on the #1 and #2 with all you have got!
• Does momentum really exist in professional football?
Yes. Not only does it exist, but it is overwhelmingly present (but all we fans already knew that,
right?).
• What key factors contribute to home team advantage in the NBA?
Shot selection strategies matter. At home, lean into the 2-pointers, because you will get the
foul calls
Brian Walsh is a Senior Technology Commercialization Manager of the UWM Research Foundation whose mission is to guide innovation, ignite startups, and foster partnerships with companies worldwide. Walsh joined the UWMRF in 2022 and has helped to grow the organization and build new programs to support UW-Milwaukee’s cutting-edge researchers and new entrepreneurs.
