1 paper
Xingcheng Fu, Xianjun Chen, Zhihao Li
Backpropagation-trained dense neural networks are powerful function approximators, but they couple learning across many parameters and can overwrite previous associations when task…