Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
June 6, 2026Journal of Optimization Differential Equations and their ApplicationsOpen Access

Combined Quasi-Newton Methods with Two-Dimensional Search for Degenerate Unconstrained Optimization in Machine Learning

View Full Paper
Ask AI
Bookmark
Share

Authors

VZViktor M. ZadachynSimon Kuznets Kharkiv National University of Economics

Discussion

Loading...

Member takes

Implication

Randomized trial demonstrates efficient optimization in machine learning, suggesting improved techniques for complex problems.

Key Points

  • This research aims to improve quasi-Newton methods for ill-conditioned and degenerate unconstrained optimization problems.
  • Developed a two-dimensional search algorithm combining quasi-Newton methods with alternative optimization methods.
  • Used spectral decomposition of the approximate Hessian for space decomposition.
  • Performed numerical experiments on standard test problems from machine learning.
  • The proposed method showed enhanced efficiency in managing spectral degeneracy compared to traditional methods.
  • Performance was benchmarked against implementations in R, Scilab, Python, and PyTorch.
  • Numerical tests indicated improved optimization results in challenging scenarios.

Cite This Study

Viktor M. Zadachyn (2025) studied this question.

synapsesocial.com/papers/6a23b8f271a5da9775e7500bhttps://doi.org/10.15421/142609
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1A Combined Quasi-Newton-Type Method Using 4th-Order Directional Derivatives for Solving Degenerate Problems of Unconstrained Optimization2024 · 1 citations
  2. 2A Quasi-Newton Primal-Dual Algorithm with Line Search2024
  3. 3Subspace Quasi-Newton Method with Gradient Approximation2024
  4. 4Combined methods for solving degenerate unconstrained optimization problems2024 · 1 citations
  5. 5Quasi-Newton method with subspace gradients2026