Autoencoder-Based Gradient Compression for Distributed Training.
Lusine AbrahamyanYiming ChenGiannis BekoulisNikos DeligiannisPublished in: EUSIPCO (2021)
Keyphrases
- feedforward neural networks
- multi agent
- cooperative
- image compression
- weighted sums
- distributed data
- test set
- distributed systems
- training samples
- distributed environment
- peer to peer
- data sets
- training examples
- training set
- computer networks
- training process
- restricted boltzmann machine
- neural network
- supervised learning
- multiresolution
- fault tolerant
- compression algorithm
- image processing
- feature selection
- compression rate