Notes or a summary of Deep Learning that i know.
- Have 7 layer (6 constructed layer, 1 layer for softmax regression)
- First 6 layers are built by DBN (Deep Belief Network)
- Deep Belief Network is using RBM (Restricted Boltzmann Machine) algorithm
- RBM constructing system that called auto encoder
- Auto encoder have responsbility to reduce number of node from layer to layer (6 first layer)
- First 6 layers not built manually but built through pre-training process with unsupervised learning. RBM algorithm is unsupervised learning category in Machine Learning.
- Jianzheng Liu, Chunlin Fang, and ChaoWu, “A Fusion Face Recognition Approach Based on 7-Layer Deep Learning Neural Network” in Journal of Electrical and Computer Engineering Vol 2016, March 2016.