Global convergence to a fixed point of the SUCPA-map

Let the SUCPA-map be f=[f1,,fK]:RKRK\mathbf{f}=[f_1,\ldots,f_K]:\mathbb{R}^K\to\mathbb{R}^K, and let ω(β[0])\omega(\boldsymbol{\beta}^{[0]}) denote the ω\omega-limit set of an initial condition β[0]\boldsymbol{\beta}^{[0]}. A point β\boldsymbol{\beta}^* is a fixed point when f(β)=β\mathbf{f}(\boldsymbol{\beta}^*)=\boldsymbol{\beta}^*. Global convergence conjecture. For each initial condition β[0]RK\boldsymbol{\beta}^{[0]}\in\mathbb{R}^K, there exists a unique fixed point βRK\boldsymbol{\beta}^*\in\mathbb{R}^K, possibly depending on β[0]\boldsymbol{\beta}^{[0]}, such that ω(β[0])={β}\omega(\boldsymbol{\beta}^{[0]})=\{\boldsymbol{\beta}^*\}. The claim concerns convergence of every SUCPA iteration to a single fixed point; its resolution is not established by the supplied text.

Sources & referencesView supporting material

Primary source

Roberta Hansen, Matias Vera, Lautaro Estienne, Luciana Ferrer and Pablo Piantanida, “On the Stability of a non-hyperbolic nonlinear map with non-bounded set of non-isolated fixed points with applications to Machine Learning”, arXiv:2401.03051 (2024).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.