Why do Larger Models Generalize Better? A Theoretical Perspective via
the XOR Problem

v1v2 (latest)

Why do Larger Models Generalize Better? A Theoretical Perspective via the XOR Problem

6 October 2018

ArXiv (abs)PDF HTML

Papers citing "Why do Larger Models Generalize Better? A Theoretical Perspective via the XOR Problem"

5 / 5 papers shown

Title
Generalization in Reinforcement Learning with Selective Noise Injection and Information BottleneckNeural Information Processing Systems (NeurIPS), 2019 Maximilian Igl K. Ciosek Yingzhen Li Sebastian Tschiatschek Cheng Zhang Sam Devlin Katja Hofmann OffRL 175 188 0 28 Oct 2019
On the Power and Limitations of Random Features for Understanding Neural Networks Gilad Yehudai Ohad Shamir MLT 282 187 0 01 Apr 2019
Gradient Descent with Early Stopping is Provably Robust to Label Noise for Overparameterized Neural Networks Mingchen Li Mahdi Soltanolkotabi Samet Oymak NoLa 394 374 0 27 Mar 2019
Overparameterized Nonlinear Learning: Gradient Descent Takes the Shortest Path? Samet Oymak Mahdi Soltanolkotabi ODL 240 184 0 25 Dec 2018
Size-Independent Sample Complexity of Neural Networks Noah Golowich Alexander Rakhlin Ohad Shamir 424 587 0 18 Dec 2017