Abstract:Let $X=(x_{ij})\in\mathbb{R}^{N\times n}$ be a rectangular random matrix with i.i.d. entries (we assume $N/n\to\mathbf{a}>1$), and denote by $\sigma_{min}(X)$ its smallest singular value. When entries have mean zero and unit second moment, the celebrated work of Bai-Yin and Tikhomirov show that $n^{-\frac{1}{2}}\sigma_{min}(X)$ converges almost surely to $\sqrt{\mathbf{a}}-1.$ However, little is known when the second moment is infinite. In this work we consider symmetric entry distributions satisfying $\mathbb{P}(|x_{ij}|>t)\sim t^{-\alpha}$ for some $\alpha\in(0,2)$, and prove that $\sigma_{min}(X)$ can be determined up to a log factor with high probability: for any $D>0$, with probability at least $1-n^{-D}$ we have $$C_1n^{\frac{1}{\alpha}}(\log n)^\frac{5(\alpha-2)}{2\alpha}\leq \sigma_{min}(X)\leq C_2n^{\frac{1}{\alpha}}(\log n)^\frac{\alpha-2}{2\alpha}$$ for some constants $C_1,C_2>0$. This appears to be the first determination of $\sigma_{min}(X)$ in the $\alpha$-stable case with a correct leading order of $n$, as previous ant-concentration arguments only yield lower bound $n^\frac{1}{2}$. The same lower bound holds for $\sigma_{min}(X+B)$ for any fixed rectangular matrix $B$ with no assumption on its operator norm. The case of diverging aspect ratio is also computed. Geometrically, the lower bound shows that the random polytope $X^*(B_1^N)$ generated by heavy-tail distributions will with very high probability contain Euclidean balls $B_2^n$ of a much larger radius compared to its Gaussian counterpart.

On random $\pm 1$ matrices: Singularity and Determinant

Precise Asymptotics for Random Matrices and Random Growth Models

Tight Asymptotic of Probability of singularity of n x n Random Matrix with Uniform Distributed \pm 1 Entries

Singularity degree of structured random matrices

On the singularity probability of random circulant Bernoulli matrices

Behavior of the minimum singular value of a random Vandermonde matrix

A Note on the Convergence Rate of the Spectral Distributions of Large Random Matrices

The singularity probability of a random symmetric matrix is exponentially small

Approximate Spielman-Teng theorems for the least singular value of random combinatorial matrices

A note on the singularity probability of random directed $d$-regular graphs

Local limits in $p$-adic random matrix theory

Universality for random matrices with an edge spectrum singularity

Non-asymptotic theory of random matrices: extreme singular values

Entropy Functions and Determinant Inequalities

On the determinants of matrices with elements from arbitrary sets

On Random Matrices Arising in Deep Neural Networks. Gaussian Case

The smallest singular value for rectangular random matrices with Lévy entries

Singularity, Wielandt's Lemma and Singular Values

The rank of a random triangular matrix over $\mathbb{F}_q$

The smallest singular value of inhomogenous random rectangular matrices

On the Characteristic Polynomial of the Eigenvalue Moduli of Random Normal Matrices