Learning Additive Noise Channels: Generalization Bounds and Algorithms

An additive noise channel is considered, in which the noise distribution is unknown and does not known to belong to any parametric family. The problem of designing a codebook and a generalized minimal distance decoder (which is parameterized by a covariance matrix) based on samples of the noise is considered. High probability generalization bounds for the error probability loss function, as well as for a hinge-type surrogate loss function are provided. A stochastic-gradient based alternating-minimization algorithm for the latter loss function is presented. Bounds on the average empirical error and generalization error are provided for a Gibbs based algorithm that gradually expurgates codewords from a large initial codebook to obtain a smaller codebook with improved error probability.