1 paper · 1 filter
Jérôme Bolte, Tam Le, Ãric Moulines +1
Motivated by the extensive application of approximate gradients in machine learning and optimization, we investigate inexact subgradient methods subject to persistent additive erro…