Kui Ren


Columbia University


Monday, February 13, 2023 - 4:00pm to 5:00pm



Zoom - https://uci.zoom.us/j/97796361534

Weighted least-squares optimization has been revisited in recent years for the computational solution of data-matching problems. Different weighting strategies have been proposed depending on the features in the solutions that one wants to promote or suppress. While the idea of using weights is quite old, there are some new understanding of it in the context of recent applications. We will review some of the recent results in this direction and highlight the impact of the weighting schemes on problems with noisy data.