Professor Daniela Calvetti will speak at the Mid-Atlantic Numerical Analysis Conference on November 15, 2019. This one-day Conference is being held at Temple University in Philadelphia, Pennsylvania. Dr. Calvetti’s talk is titled “The Unreasonable Effectiveness of Numerical Linear Algebra in Bayesian Inverse Problems.”
Inverse problems deal with recovering unknown causes from incomplete, noisy indirect observations of their effect. This task may be very difficult even when the observation model is linear, especially if the forward map is very smoothing. In this talk we show how, recently, numerical linear algebra and Bayesian inference have joined forces to provide amazing solutions to a wide range of challenging inverse problems.