Non-Gaussian Component Analysis via Lattice Basis Reduction
This resolves a theoretical gap in distribution learning for discrete settings, offering an efficient solution where tradeoffs were previously suspected.
The paper tackles the problem of Non-Gaussian Component Analysis (NGCA) for discrete or nearly discrete distributions, where prior work suggested an information-computation tradeoff, and provides a sample and computationally efficient algorithm using lattice basis reduction, achieving efficient approximation of the hidden direction.
Non-Gaussian Component Analysis (NGCA) is the following distribution learning problem: Given i.i.d. samples from a distribution on $\mathbb{R}^d$ that is non-gaussian in a hidden direction $v$ and an independent standard Gaussian in the orthogonal directions, the goal is to approximate the hidden direction $v$. Prior work \cite{DKS17-sq} provided formal evidence for the existence of an information-computation tradeoff for NGCA under appropriate moment-matching conditions on the univariate non-gaussian distribution $A$. The latter result does not apply when the distribution $A$ is discrete. A natural question is whether information-computation tradeoffs persist in this setting. In this paper, we answer this question in the negative by obtaining a sample and computationally efficient algorithm for NGCA in the regime that $A$ is discrete or nearly discrete, in a well-defined technical sense. The key tool leveraged in our algorithm is the LLL method \cite{LLL82} for lattice basis reduction.