Automatic node selection for Deep Neural Networks using Group Lasso regularization.
Tsubasa OchiaiShigeki MatsudaHideyuki WatanabeShigeru KatagiriPublished in: ICASSP (2017)
Keyphrases
- group lasso
- neural network
- regression model
- variable selection
- mixed norm
- efficient optimization
- regularization term
- contingency tables
- node selection
- multiple kernel learning
- multi task
- low rank
- multi task learning
- data sets
- multi label learning
- convex optimization
- energy functional
- active learning
- image segmentation
- keyword search
- multi label
- high dimensional
- machine learning
- semi supervised
- wireless sensor networks