MLP training dynamics follow saddle-organized plateaus and near-optima before necessarily settling into an overfitting attractor on finite noisy datasets.
Title resolution pending
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.LG 2verdicts
UNVERDICTED 2representative citing papers
A method that alternates gradient steps on a neural network backbone with closed-form optimal updates to the final linear layer under squared loss, including an SGD adaptation and NTK-regime convergence analysis.
citing papers explorer
-
Plateaus, Optima, and Overfitting in Multi-Layer Perceptrons: A Saddle-Saddle-Attractor Scenario
MLP training dynamics follow saddle-organized plateaus and near-optima before necessarily settling into an overfitting attractor on finite noisy datasets.
-
Closed-Form Last Layer Optimization
A method that alternates gradient steps on a neural network backbone with closed-form optimal updates to the final linear layer under squared loss, including an SGD adaptation and NTK-regime convergence analysis.