Two-argument activation functions learn soft XOR operations like cortical neurons

Open in new window