Floating-Point Neural Networks are Provably Robust Universal Approximators.

Geonho Hwang, Wonyeol Lee, Yeachan Park, Sejun Park, Feras Saad

Browse the full CAV paper archive.