Online Learning and Information Exponents: On The Importance of Batch size, and Time/Complexity Tradeoffs

Documenti analoghi

• Repetita Iuvant: Data Repetition Allows SGD to Learn High-Dimensional Multi-Index Functions
  di: Arnaboldi, Luca, et al.
  Pubblicazione: (2024)
• The Benefits of Reusing Batches for Gradient Descent in Two-Layer Networks: Breaking the Curse of Information and Leap Exponents
  di: Dandi, Yatin, et al.
  Pubblicazione: (2024)
• How Two-Layer Neural Networks Learn, One (Giant) Step at a Time
  di: Dandi, Yatin, et al.
  Pubblicazione: (2023)
• Deep Learning of Compositional Targets with Hierarchical Spectral Methods
  di: Tabanelli, Hugo, et al.
  Pubblicazione: (2026)
• The Computational Advantage of Depth: Learning High-Dimensional Hierarchical Functions with Gradient Descent
  di: Dandi, Yatin, et al.
  Pubblicazione: (2025)