Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs).
Djork-Arné ClevertThomas UnterthinerSepp HochreiterPublished in: ICLR (Poster) (2016)
Keyphrases
- learning process
- learning algorithm
- online learning
- recurrent networks
- deep learning
- quality of service
- network architecture
- deep architectures
- connectionist networks
- spiking neural networks
- fully connected
- learning scheme
- network model
- incremental learning
- recurrent neural networks
- learning problems
- computationally efficient
- unsupervised learning
- supervised learning