arXiv subjects
Publications and source records attributed to M. Rudelson.
Let $A$ be an $n \times n$ random matrix with independent identically distributed non-constant subgaussian entries. Then for any $k \le c \sqrt{n}$, \[ \text{rank}(A) \ge n-k \] with probability at least $1-\exp(-c'kn)$.
Let $K$ be an $n$-dimensional convex body. Define the difference body by $$ K-K= \{x-y \mid x,y \in K \}. $$ We estimate the volume of the section of $K-K$ by a linear subspace $F$ via the maximal volume of sections of $K$ parallel to $F$. We prove that for any $m$-dimensional subspace $F$ there exists $x \in R^n$, such that $$ vol ((K-K) \cap F) \le C^m (\min (n/m, \sqrt{m}))^m \cdot vol (K \cap (F+x)), $$ for some absolute constant $C$. We show that for small dimensions of $F$ this estimate is exact up to a multiplicative constant.
Let $K, D$ be $n$-dimensional convex bodes. Define the distance between $K$ and $D$ as $$ d(K,D) = \inf \{λ| T K \subset D+x \subset λ\cdot TK \}, $$ where the infimum is taken over all $x \in R^n$ and all invertible linear operators $T$. Assume that 0 is an interior point of $K$ and define $$ M(K) =\int_{S^{n-1}} \| ω\|_K d μ(ω), $$ where $μ$ is the uniform measure on the sphere. Let $K^{\circ}$ be the polar body of $K$. We use the difference body estimate to prove that $K$ can be embedded into $R^n$ so that $$ M(K) \cdot M(K^{\circ}) \le C n^{1/3} \log^a n $$ for some absolute constants $C$ and $a$. We apply this result to show that the distance between two $n$-dimensional convex bodies does not exceed $n^{4/3}$ up to a logarithmic factor.