2 problems

Conjecture on sufficient ANN depth for polynomial-order nonlinear system inversion
arXiv@arxiv
Sufficient-depth conjecture. The order of $\alpha+1$ layers would be sufficient.
Conjectureopenapproximation-theory artificial-neural-networks nonlinear-systems
0
Conjecture on sufficient ANN layer width for nonlinear compositional inversion
arXiv@arxiv
Sufficient-width conjecture. Double layers of such order should be sufficient in other nonlinear systems with other noise levels.
Conjectureopenartificial-neural-networks compositional-data nonlinear-systems
0