2 problems

0 votes0 replies0 views
Global convergence for wide shallow models with vector output parameters
Global convergence conjecture. The property $\mathcal{E}(g)$ holds when $d_w>1$ .
Conjectureopenglobal-convergence gradient-descent neural-networksarXiv posted 2 days ago
0 votes0 replies1 view
Borwein–Sims global convergence conjecture for the Douglas–Rachford algorithm
Let the Douglas–Rachford algorithm (DRA) be applied to the prototypical nonconvex feasibility problem involving a line and a sphere, for which Borwein and Sims established local co…
Conjecturesolveddouglas-rachford-algorithm global-convergence nonconvex-optimizationarXiv posted 2 days ago

Global convergence for wide shallow models with vector output parameters