A Provable Generalized Tensor Spectral Method for Uniform Hypergraph Partitioning

Matrix spectral methods play an important role in statistics and machine learning, and most often the word `matrix’ is dropped as, by default, one assumes that similarities or affinities are measured