add attention method
(1) Densely Connected CNN with Multi-scale Feature Attention for Text Classification
Add attention on the last layer before classification.
(2) Attention-Mechanism-Containing Neural Networks for High-Resolution Remote Sensing Image Classification
(3) Learning Multi-Attention Convolutional Neural Network for Fine-Grained Image Recognition
(4) LEARN TO PAY ATTENTION
(5) Weighted Channel Dropout for Regularization of Deep Convolutional Neural Network