SGD Finds then Tunes Features in Two-Layer Neural Networks with near-Optimal Sample Complexity: A Case Study in the XOR problem.

Margalit Glasgow

Published in: ICLR (2024)

Keyphrases

sample complexity
neural network
feature extraction
upper bound
theoretical analysis
feature space
feature set
feature vectors
active learning
training examples
learning problems
vc dimension
pac learning
prior knowledge
small number