Elementary functions such as sin or exp may naively be considered as good generators of random bits: the bit-runs output by these functions axe believed to be statistically random most of the time. Here we investigate their computational hardness: given a part of the binary expansion of exp x, can one recover x? We describe a heuristic technique to address this type of questions. It relies upon Coppersmith's heuristic technique - itself based on lattice reduction - for finding the small roots of multivariate polynomials modulo an integer. For our needs, we improve the lattice construction step of Coppersmith's method: we describe a way to find a subset of a set of vectors that decreases the Minkowski theorem bound, in a rather general setup including Coppersmith-type lattices. © Springer-Verlag Berlin Heidelberg 2006.
CITATION STYLE
Stehlé, D. (2006). On the randomness of bits generated by sufficiently smooth functions. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 4076 LNCS, pp. 257–274). Springer Verlag. https://doi.org/10.1007/11792086_19
Mendeley helps you to discover research relevant for your work.