DNN에서의 vanishing gradient 및 degradation problem을 해결하기 위한 방안으로 Skip Connection을 제시하였다.
1. Architecture
1.1. Fully Connected Layer
1.1.1. ReLU
1.2. Skip Connection
References
[1] He, Kaiming, et al. “Deep residual learning for image recognition.” Proceedings of the IEEE conference on computer vision and pattern recognition. 2016.