Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper analyzes the asymptotic behavior of functionals of Gaussian outputs from infinitely-wide random neural networks on the d-dimensional sphere as network depth increases. It identifies three distinct limiting regimes for these functionals: convergence to a functional of a limiting Gaussian field, convergence to a Gaussian distribution, and convergence to a distribution in the Qth Wiener chaos. The key result is that the specific limiting regime is governed by the fixed-point structure of the covariance function's iterative operator, revealing how network depth impacts the statistical properties of network outputs.
Random neural networks exhibit surprisingly distinct statistical behaviors depending on the fixed points of their covariance functions as depth increases, revealing a phase transition-like phenomenon.
We establish central and non-central limit theorems for sequences of functionals of the Gaussian output of an infinitely-wide random neural network on the d-dimensional sphere . We show that the asymptotic behaviour of these functionals as the depth of the network increases depends crucially on the fixed points of the covariance function, resulting in three distinct limiting regimes: convergence to the same functional of a limiting Gaussian field, convergence to a Gaussian distribution, convergence to a distribution in the Qth Wiener chaos. Our proofs exploit tools that are now classical (Hermite expansions, Diagram Formula, Stein-Malliavin techniques), but also ideas which have never been used in similar contexts: in particular, the asymptotic behaviour is determined by the fixed-point structure of the iterative operator associated with the covariance, whose nature and stability governs the different limiting regimes.