Generalization analysis with deep ReLU networks for metric and similarity learning

no code implementations • 10 May 2024 • Junyu Zhou, Puyu Wang, Ding-Xuan Zhou

Specifically, by deriving the explicit form of the true metric for metric and similarity learning with the hinge loss, we construct a structured deep ReLU neural network as an approximation of the true metric, whose approximation ability relies on the network complexity.

Optimal Estimates for Pairwise Learning with Deep ReLU Networks

no code implementations • 31 May 2023 • Junyu Zhou, Shuo Huang, Han Feng, Ding-Xuan Zhou

Pairwise learning refers to learning tasks where a loss takes a pair of samples into consideration.