SIAM awarded the 2017 SIAM Student Paper Prize to Bamdad Hosseini for his paper, “Well-Posed Bayesian Inverse Problems: Priors with Exponential Tails,” co-authored with Nilima Nigam of Simon Fraser University and published in SIAM/ASA Journal on Uncertainty Quantification in 2017. Hosseini received the award and presented his winning paper at the SIAM Annual Meeting, held July 10-14, 2017 in Pittsburgh, Pennsylvania.
The SIAM Student Paper Prize is awarded annually to the student authors of the most outstanding papers nominated for the prize. The 2017 Student Paper Prize recognized outstanding scholarship by students in applied mathematics and computing as evidenced in a paper submitted for publication in a peer-reviewed journal. The awards are based solely on the merit and content of the student’s contribution to the paper. Up to three awards are given each year.
Bamdad Hosseini is currently a postdoctoral scholar in the Department of Computing and Mathematical Sciences at California Institute of Technology, under the supervision of Andrew Stuart. He received his PhD in 2017 in applied and computational mathematics from Simon Fraser University, where his supervisors were John Stockie and Nilima Nigam. He earned his MS at Simon Fraser University in 2013 and his BS in mechanical engineering in 2011 at Sharif University of Technology, Tehran, Iran. His research centers on the analysis, development, and application of methods for estimating parameters and quantifying uncertainty. His particular interest is in the Bayesian approach for solution of inverse problems.
Q: Why are you excited about winning the prize?
A: SIAM is the most prominent professional society for applied mathematicians in North America. As a graduate student I learned a lot from papers and books that were published by SIAM and I continue to use such publications in my work. It is a great honor to be recognized by such an amazing community.
Q: What does your research mean to the public?
A: Broadly speaking, I am interested in the inference of unknowns from noisy data and estimation of the uncertainties that are associated with these unknowns. Nowadays fast computers and large amounts of data are accessible to most of us but we need mathematical models and algorithms that allow us to extract useful information from these data sets. I like to analyze these models and algorithms to obtain a better understanding of their behavior and possibly improve them.
Q: Could you tell us a bit about the research that won you the prize?
A: A common approach for solution of under-determined inverse problems is to find the minimizer of a penalized least squares functional. In practice, we often choose convex penalization terms such as the well-known Tikhonov regularizer. This penalized least squares approach has a probabilistic interpretation using Bayes' rule that allows us to combine physical measurements and our prior knowledge of the solution to solve an inverse problem. Such an inverse problem is called a 'Bayesian inverse problem' and its solution is an entire probability distribution on the unknown rather than a point estimator. The main contribution of our paper was to analyze the well-posedness and consistent approximation of infinite dimensional Bayesian inverse problems with convex prior distributions that are analogous to variational inverse problems with convex regularization terms.
Q: What does being a SIAM member mean to you?
A: Most importantly for me, being a member of SIAM means having access to the cutting edge of applied mathematics. As a SIAM member it is much easier for me to network with researchers in my field and find new collaborations. I should also mention the multitude of professional development opportunities provided by SIAM for students, such as the SIAM student chapters and the career panels during the SIAM conferences, from which I have benefited.