The large $N$ factorization does not hold for arbitrary multi-trace observables in random tensors
Abstract
We consider real tensors of order , that is -dimensional arrays of real numbers , where each index can take values. The tensor entries have no symmetry properties under permutations of the indices. The invariant polynomials built out of the tensor entries are called trace invariants. We prove that for a Gaussian random tensor with indices (that is such that the entries are independent identically distributed Gaussian random variables) the cumulant, or connected expectation, of a product of trace invariants is not always suppressed in scaling in with respect to the product of the expectations of the individual invariants. Said otherwise, not all the multi-trace expectations factor at large in terms of the single-trace ones and the Gaussian scaling is not subadditive on the connected components. This is in stark contrast to the case of random matrices in which the multi-trace expectations always factor at large . The best one can do for is to identify restricted families of invariants for which the large factorization holds and we check that this indeed happens when restricting to the family of melonic observables, the dominant family in the large limit.
Cite
@article{arxiv.2506.15362,
title = {The large $N$ factorization does not hold for arbitrary multi-trace observables in random tensors},
author = {Razvan Gurau and Felix Joos and Benjamin Sudakov},
journal= {arXiv preprint arXiv:2506.15362},
year = {2025}
}