Fetching the paper…

Enhancing approximation abilities of neural networks by training derivatives · Around