Let A be an n by n random matrix with independent identically distributed entries, such that probability that an entry is non-zero is sublogarithmic in n. In particular, this implies that with large probability some of rows and columns of A are zero. In this talk, I will consider the problem of estimating the smallest singular value of shifted matrix A-zI, for a non-zero complex number z. This estimate is crucial in studying the limiting spectral distribution for a sequence of very sparse random matrices.
Based on a joint work with Mark Rudelson.
Back to Quantitative Linear Algebra