A Count-sketch to Reduce Memory Consumption when Training a Model with Gradient Descent.
Wissam SibliniFrank MeyerPascale KuntzPublished in: IJCNN (2019)
Keyphrases
- objective function
- cost function
- computational model
- neural network
- formal model
- probabilistic model
- management system
- statistical model
- probability distribution
- em algorithm
- mathematical model
- piecewise constant
- neural network model
- test data
- test set
- parameter estimation
- theoretical analysis
- maximum likelihood
- semi supervised
- multi agent
- similarity measure
- feature selection