1 paper
Xiaoguang Chang, Teng Wang, Changyin Sun
The simplicity Bias (SB) of neural networks, i.e.\ their tendency to represent simple functions, is a key factor in their generalization capabilities. Recent studies show that an e…