{"id":"arxiv/2603.01342","url":"https://arxiv.org/abs/2603.01342","tags":["math.PR","math-ph","quant-ph"],"score":0,"title":"A moment-based approach to the injective norm of random tensors","domain":"math.PR","source":"arxiv","authors":["Stephane Dartois","Benjamin McKenna"],"reviews":0,"subject":"math.PR","summary":"In this paper, we present a technically simple method to establish upper bounds on the expected injective norm of real and complex random tensors. Our approach is somewhat analogous to the moment method in random matrix theory, and is based on a deterministic upper bound on the injective norm of a tensor which might be of independent interest. Compared to previous approaches to these problems (spin-glass methods, epsilon-net techniques, Sudakov-Fernique arguments, and PAC-Bayesian proofs), our method has the benefit of being nonasymptotic, relatively elementary, and applicable to…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-03-02","reactions":{}}