Sparse Networks from Scratch: Faster Training without Losing Performance.
Tim DettmersLuke ZettlemoyerPublished in: CoRR (2019)
Keyphrases
- echo state networks
- network topologies
- high dimensional
- recurrent networks
- memory efficient
- network structure
- genetic algorithm
- avoid overfitting
- training examples
- training process
- complex networks
- compressive sensing
- training phase
- training algorithm
- linear combination
- supervised learning
- small number
- training set
- social networks