We consider a class of sparse random matrices of the form $A_n =(��_{i,j}��_{i,j})_{i,j=1}^n$, where $\{��_{i,j}\}$ are i.i.d.~centered random variables, and $\{��_{i,j}\}$ are i.i.d.~Bernoulli random variables taking value $1$ with probability $p_n$, and prove a quantitative estimate on the smallest singular value for $p_n = ��(\frac{\log n}{n})$, under a suitable assumption on the spectral norm of the matrices. This establishes the invertibility of a large class of sparse matrices. For $p_n =��( n^{-��})$ with some $��\in (0,1)$, we deduce that the condition number of $A_n$ is of order $n$ with probability tending to one under the optimal moment assumption on $\{��_{i,j}\}$. This in particular, extends a conjecture of von Neumann about the condition number to sparse random matrices with heavy-tailed entries. In the case that the random variables $\{��_{i,j}\}$ are i.i.d.~sub-Gaussian, we further show that a sparse random matrix is singular with probability at most $\exp(-c n p_n)$ whenever $p_n$ is above the critical threshold $p_n = ��(\frac{ \log n}{n})$. The results also extend to the case when $\{��_{i,j}\}$ have a non-zero mean. We further find quantitative estimates on the smallest singular value of the adjacency matrix of a directed Erd��s-R��yni graph whenever its edge connectivity probability is above the critical threshold $��(\frac{\log n}{n})$.

46 pages, minor changes in V3. To appear in Advances in Mathematics