Paper Reading:Learning Transferable Architectures for Scalable Image Recognition

Learning Transferable Architectures for Scalable Image Recognition

Motivation

NAS直接搜全部结构方法太慢，设计一个搜索空间较小的方法。即类似于LSTM中的子结构，重复堆叠这些子结构，从而得到较深的网络。

Architecture

对于两种不同的数据集，大致的设计了两种不同的网络结构，

确定好整体结构后，按照以下步骤

• Step 1. Select a hidden state from hi, hi−1 or from the set of hidden states created in previous blocks.

• Step 2. Select a second hidden state from the same options as in Step 1.

• Step 3. Select an operation to apply to the hidden state selected in Step 1.

• Step 4. Select an operation to apply to the hidden state selected in Step 2.

• Step 5. Select a method to combine the outputs of Step 3 and 4 to create a new hidden state.
  

第三步和第四步的操作合集如下，

第五步的操作合集如下，

(1) element-wise addition be- tween two hidden states or

(2) concatenation between two hidden states along the filter dimension

Experiment

最终得到的网络结果如下：

在分类上的结果，