“…Here, the target dimension k is much smaller than the original dimension d. The name random projections was coined after the first construction by Johnson and Lindenstrauss in [1] who showed that such mappings exist for k ∈ O(log(1/δ)/ε 2 ). Other constructions of random projection matrices have been discovered since [2,3,4,5,6]. Their properties make random projections a key player in rank-k approximation algorithms [7,8,9,10,11,12,13,14], other algorithms in numerical linear algebra [15,16,17], compressed sensing [18,19,20], and various other applications, e.g, [21,22].…”