Increasing Network Size and Training Throughput of FPGA Restricted Boltzmann Machines Using Dropout.
Jiang SuDavid B. ThomasPeter Y. K. CheungPublished in: FCCM (2016)
Keyphrases
- restricted boltzmann machine
- network size
- contrastive divergence
- deep learning
- learning algorithm
- communication cost
- conditional random fields
- activation function
- probabilistic graphical models
- deep belief networks
- network structure
- hidden layer
- training phase
- semi supervised
- optical flow
- hardware implementation
- training set
- supervised learning
- training examples
- neural network