A Rigorous Analysis of Linsker-type Hebbian Learning