ProxyConvexity: AUnifiedFramework fortheAnalysisofNeuralNetworks TrainedbyGradientDescent

Feb-8-2026, 09:26:56 GMT–Neural Information Processing Systems

Weintroduce thenotions of proxy convexity and proxy Polyak-Lojasiewicz (PL) inequalities, which are satisfied iftheoriginal objectivefunction induces aproxy objectivefunction that is implicitly minimized when using gradient methods.

artificial intelligence, inequality, machine learning, (14 more...)

Neural Information Processing Systems

Feb-8-2026, 09:26:56 GMT

Conferences PDF

Add feedback

Technology:
- Information Technology > Artificial Intelligence > Machine Learning
  - Neural Networks (0.31)
  - Statistical Learning (0.48)

Duplicate Docs Excel Report

Title
42299f06ee419aa5d9d07798b56779e2-Paper.pdf

Similar Docs Excel Report more

Title	Similarity	Source
None found