Learning the Group Structure of Deep Neural Networks with an Expectation Maximization Method.
Subin YiJaesik ChoiPublished in: ICDM Workshops (2018)
Keyphrases
- em algorithm
- neural network
- expectation maximization
- high accuracy
- prior knowledge
- dynamic programming
- neural nets
- probabilistic model
- detection method
- training process
- significant improvement
- cost function
- preprocessing
- pattern recognition
- unsupervised learning
- learning mechanism
- graphical models
- similarity measure
- feed forward neural networks
- recurrent neural networks
- gaussian mixture model
- segmentation method
- parameter estimation
- supervised learning
- k means
- computational complexity
- learning algorithm