EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 4.0

Basins of attraction in neural network training: a Julia-Fatou topological view of loss landscapes

Busayo Samuel, Moses Obinna Francis

Annals of Mathematics and Computer Science · 20.03.2026 · 33

DOI: 10.56947/amcs.v33.737 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Annals of Mathematics and Computer Science, online erschienen am 20.03.2026. Verfasst von Busayo Samuel und 1 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

We study gradient-based training in deep learning from a complex-dynamical perspective. Under a local analytic continuation assumption, the training update is modeled as iteration of a holomorphic map, which organizes the loss landscape into stable regions and sensitive boundary sets. Using fixed-point stability ideas, we derive explicit local criteria distinguishing attraction from repulsion and obtain a linear convergence rate inside attracting neighborhoods. We also extend the local stability analysis to higher-dimensional complex parameter spaces via a spectral condition. At the global level, we prove three complementary results: an escape-radius condition that forces divergence for polynomial-gradient surrogates, persistence of attracting fixed points under small learning-rate perturbations, and an instability principle for rational maps based on the density of repelling periodic points on the Julia set. These results link step-size sensitivity and initialization dependence to basin geometry and support visualization-driven diagnostics for stability in optimization.

Zitieren

APA

Samuel, B., & Francis, M. O. (2026). Basins of attraction in neural network training: a Julia-Fatou topological view of loss landscapes. Annals of Mathematics and Computer Science, 33. https://doi.org/10.56947/amcs.v33.737

Vancouver

Samuel B, Francis MO. Basins of attraction in neural network training: a Julia-Fatou topological view of loss landscapes. Annals of Mathematics and Computer Science. 2026;33. doi:10.56947/amcs.v33.737

BibTeX
@article{Samuel2026,
  author = {Samuel, Busayo and Francis, Moses Obinna},
  title = {Basins of attraction in neural network training: a Julia-Fatou topological view of loss landscapes},
  journal = {Annals of Mathematics and Computer Science},
  year = {2026},
  volume = {33},
  doi = {10.56947/amcs.v33.737}
}
Kostenlos starten →