Open problem: Improper learning of mixtures of Gaussians
[edit]
Proceedings of the 31st Conference On Learning Theory, PMLR 75:33993402, 2018.
Abstract
We ask whether there exists an efficient unsupervised learning algorithm for mixture of Gaussians in the overcomplete case (number of mixtures is larger than the dimension). The notion of learning is taken to be worstcase compressionbased, to allow for improper learning.
Related Material


