Improvement of learning for CNN with ReLU activation by sparse regularization.

Hidenori Ide, Takio Kurita

Browse the full IJCNN paper archive.