| 
Algorithm 1 GA-based NAS Algorithm | 
- 
1:
  Input: the number of generations T, the number of individuals in each generation N,  
the dataset D; 
 
- 
2:
  Initialization: generating the initial individuals randomly; 
 
- 
3:
  For 1,2,3, …, T
do; 
 
- 
4:
  For 1,2,3, …, N do; 
 
- 
5:
  Evaluation: neural network architecture (individual) training for D_train; 
 
- 
6:
  Evaluate the performance of the neural network architecture for D_test; 
 
- 
7:
  end for 
 
- 
8:
  Selection: selecting the most suitable individuals for reproduction; 
 
- 
9:
  Crossover and Mutation: reproducing new individuals;  
 
- 
10:
end for 
 
- 
11:
Output: individuals in the last generation. 
 
  |