Showing cs.LGShow all
2 papers · 1 filter
cs.LG2023
Neural Characteristic Activation Analysis and Geometric Parameterization for ReLU Networks
Wenlin Chen, Hong Ge
We introduce a novel approach for analyzing the training dynamics of ReLU networks by examining the characteristic activation boundaries of individual ReLU neurons. Our proposed an…
cs.LG2022
Understanding Sparse Feature Updates in Deep Networks using Iterative Linearisation
Adrian Goldwaser, Hong Ge
Larger and deeper networks generalise well despite their increased capacity to overfit. Understanding why this happens is theoretically and practically important. One recent approa…