More is Better in Modern Machine Learning: when Infinite Overparameterization is Optimal and Overfitting is Obligatory

Documenti analoghi

• The lazy (NTK) and rich ($μ$P) regimes: a gentle tutorial
  di: Karkada, Dhruva
  Pubblicazione: (2024)
• Predicting kernel regression learning curves from only raw data statistics
  di: Karkada, Dhruva, et al.
  Pubblicazione: (2025)
• Closed-Form Training Dynamics Reveal Learned Features and Linear Structure in Word2Vec-like Models
  di: Karkada, Dhruva, et al.
  Pubblicazione: (2025)
• Benign, Tempered, or Catastrophic: A Taxonomy of Overfitting
  di: Mallinar, Neil, et al.
  Pubblicazione: (2022)
• Alternating Gradient Flows: A Theory of Feature Learning in Two-layer Neural Networks
  di: Kunin, Daniel, et al.
  Pubblicazione: (2025)