Gradient descent provably escapes saddle points in the training of shallow ReLU networks

Documenti analoghi

• Non-convergence to global minimizers in data driven supervised deep learning: Adam and stochastic gradient descent optimization provably fail to converge to global minimizers in the training of deep neural networks with ReLU activation
  di: Do, Thang, et al.
  Pubblicazione: (2024)
• On the existence of minimizers in shallow residual ReLU neural network optimization landscapes
  di: Dereich, Steffen, et al.
  Pubblicazione: (2023)
• Is ReLU Adversarially Robust?
  di: Sooksatra, Korn, et al.
  Pubblicazione: (2024)
• The Inhibitor: ReLU and Addition-Based Attention for Efficient Transformers under Fully Homomorphic Encryption on the Torus
  di: Brännvall, Rickard, et al.
  Pubblicazione: (2023)
• The Non-Linearity Perturbation Threshold: Width Scaling and Landscape Bifurcations in Deep Learning
  di: Alexander, Michael
  Pubblicazione: (2026)