Randomized coordinate gradient descent almost surely escapes strict saddle points

Documenti analoghi

• Method of Successive Approximations for Stochastic Optimal Control: Contractivity and Convergence
  di: Taoufik, Safouane, et al.
  Pubblicazione: (2024)
• Avoidance of non-strict saddle points by blow-up
  di: Achour, El Mehdi, et al.
  Pubblicazione: (2025)
• Gradient descent avoids strict saddles with a simple line-search method too
  di: Muşat, Andreea-Alexandra, et al.
  Pubblicazione: (2025)
• Error analysis for stochastic gradient optimization schemes using modified equations
  di: Bréhier, Charles-Edouard, et al.
  Pubblicazione: (2024)
• Interpolatory model reduction of dynamical systems with root mean squared error
  di: Reiter, Sean, et al.
  Pubblicazione: (2024)