David Banks, Duke University
On August 31, after 19 years, the Statistical and Applied Mathematical Sciences Institute (SAMSI) closed its doors. I am honored to have succeeded Richard Smith and Jim Berger as its director.
SAMSI did a lot of good for statistics (and applied mathematics). At this occasion, it seems important to note and celebrate its successes. Its postdoctoral fellows program was a ladder for many successful new careers. SAMSI also helped build research foci and professional networks for junior and mid-career scientists, inside and outside of academia. Its programs helped direct the attention of our field toward important problems, and thus it shaped our shared intellectual trajectory during and beyond its lifetime. The research partnerships forged at SAMSI are durable and shall continue for decades to come.
The closing ceremony held August 21 was an opportunity to reflect on SAMSI’s influence. It was seminal in building the modern field of uncertainty quantification, which is now an integral part of the critical discipline of computer experimentation. Uncertainty quantification has been used to approximate pyroclastic flows, animal movement in agent-based models, and weather modeling. It represents a true synthesis of statistics and applied mathematics. SAMSI held two year-long programs adjacent to that topic.
SAMSI also had multiple programs on astrostatistics. That area was pioneered by Jogesh Babu and Eric Feigelson (an astronomer) at Penn State, but SAMSI gave them a place to stand from which they moved the world. There is now an active Astrostatistics Interest Group in the ASA, and many people are engaged with the data-rich problems driving modern work. New telescopes and sensors have turned this into a big data showcase for interdisciplinary research.
Risk analysis existed long before SAMSI was born, but SAMSI touched on the field in distinctive ways. The recent program on games, decisions, risk, and reliability—led by Ernest Fokoué (University of Rochester), David Ríos Insua (ICMAT), Fabrizio Ruggeri (Italian Research Council in Milan), and Refik Soyer (The George Washington University)—took on a number of novel risk topics, including the measurement of the safety of autonomous vehicles and cyberinsurance. And a previous SAMSI program focusing on national security led to the invention of adversarial risk analysis. Google Scholar lists more than 60 papers with that phrase in their titles.
A fourth area in which SAMSI played a significant role was environmental and climate studies. Bo Li at the University of Illinois Urbana-Champaign, Murali Haran at Penn State, Doug Nychka at the Colorado School of Mines, Amy Braverman at the Jet Propulsion Lab, Lenny Smith at Virginia Tech, and Veronica Berrocal at the University of California, Irvine were all key players in that effort. Significant topics included measurement of sea ice, fate and transport models for pollutants, and mathematical models for global climate change.
SAMSI had many other important programs. In 2015–2016, there was a program on forensic statistics, which was one of several responses by our community to the National Research Council’s report decrying the problematic reliability of forensic science. And there were programs on topological data analysis, computational advertising, neuroscience and brain connectomics, several flavors of social science, and the now quaintly named program on data mining.
So, we should come to praise SAMSI, not bury it. It invigorated our field, created new opportunities and challenges, launched and advanced careers, and put our best foot forward in the scientific world.
Leave a Reply