Ensemble-Compression: A New Method for Parallel Training of Deep Neural Networks.
Shizhao SunWei ChenJiang BianXiaoguang LiuTie-Yan LiuPublished in: ECML/PKDD (1) (2017)
Keyphrases
- neural network
- detection method
- training process
- training algorithm
- pairwise
- similarity measure
- clustering method
- feed forward neural networks
- training phase
- training set
- significant improvement
- fuzzy logic
- high accuracy
- test data
- model selection
- recurrent neural networks
- classification algorithm
- segmentation method
- training samples
- preprocessing
- objective function
- feature selection
- machine learning