"Oddball SGD": Novelty Driven Stochastic Gradient Descent for Training Deep Neural Networks

Devs

"Oddball SGD": Novelty Driven Stochastic Gradient Descent for Training Deep Neural Networks | Read Paper on Bytez