Login / Signup

Limitations of Information-Theoretic Generalization Bounds for Gradient Descent Methods in Stochastic Convex Optimization.

Mahdi HaghifamBorja Rodríguez GálvezRagnar ThobabenMikael SkoglundDaniel M. RoyGintare Karolina Dziugaite
Published in: CoRR (2022)
Keyphrases
  • information theoretic
  • convex optimization
  • information theory
  • mutual information
  • low rank
  • learning theory
  • machine learning
  • feature selection
  • similarity measure
  • high dimensional
  • convex optimization problems