2000 character limit reached
Fast binary embeddings with Gaussian circulant matrices: improved bounds (1608.06498v2)
Published 23 Aug 2016 in cs.IT, cs.DS, and math.IT
Abstract: We consider the problem of encoding a finite set of vectors into a small number of bits while approximately retaining information on the angular distances between the vectors. By deriving improved variance bounds related to binary Gaussian circulant embeddings, we largely fix a gap in the proof of the best known fast binary embedding method. Our bounds also show that well-spreadness assumptions on the data vectors, which were needed in earlier work on variance bounds, are unnecessary. In addition, we propose a new binary embedding with a faster running time on sparse data.