Why Not to Use Zero Imputation? Correcting Sparsity Bias in Training Neural Networks.
Joonyoung YiJuhyuk LeeKwang Joon KimSung Ju HwangEunho YangPublished in: ICLR (2020)
Keyphrases
- neural network
- training process
- training algorithm
- feed forward neural networks
- feedforward neural networks
- multi layer perceptron
- training patterns
- back propagation
- backpropagation algorithm
- neural network structure
- artificial neural networks
- neural network training
- radial basis function network
- pattern recognition
- missing values
- multilayer perceptron
- error back propagation
- training phase
- genetic algorithm
- multi layer
- neural network model
- fault diagnosis
- sparse representation
- test set
- training set
- training data
- activation function
- sufficient conditions
- supervised learning
- recurrent networks
- nearest neighbor
- reinforcement learning
- data imputation