About this Event
Abstract: In this dissertation, we view matrix frames as representations of quivers and study them within the general framework of quiver invariant theory. We are particularly interested in radial isotropic and Parseval matrix frames. Using methods from quiver invariant theory, we first prove a far-reaching generalization of Franck Barthe's Theorem on vectors in radial isotropic position to the case of matrix frames. With this tool at our disposal, we generalize the Paulsen problem from frames (of vectors) to frames of matrices of arbitrary rank and size extending Hamilton-Moitra's upper bound. Specifically, we show that for any given \epsilon-nearly equal-norm Parseval frame F of n matrices with d rows there exists an equal-norm Parseval frame W of n matrices with d rows such that the squared distance between F and W is bounded above by 46\epsilon d^2. Finally, we address the constructive aspects of transforming a matrix frame into radial isotropic position which opens the possibility to approach the problem algorithmically.
0 people are interested in this event