NeuMiss networks: differentiable programming for supervised learning with missing values.
Marine Le MorvanJulie JosseThomas MoreauErwan ScornetGaël VaroquauxPublished in: NeurIPS (2020)
Keyphrases
- missing values
- supervised learning
- missing data
- incomplete data
- missing data imputation
- incomplete data sets
- high dimensional data
- unsupervised learning
- data imputation
- exponential family
- training set
- neural network
- missing information
- semi supervised
- data streams
- training data
- database
- cross validation
- loss function
- data matrix
- feature extraction
- machine learning