Memory in Echo State Networks and the Controllability Matrix Rank.
Brian WhiteakerPeter GerstoftPublished in: ICASSP (2022)
Keyphrases
- echo state networks
- recurrent neural networks
- singular values
- nuclear norm minimization
- memory requirements
- singular value decomposition
- memory usage
- random access
- memory space
- low rank
- linear algebra
- computing power
- partial observability
- memory size
- adjacency matrix
- limited memory
- database
- support vector
- learning algorithm
- genetic algorithm
- machine learning
- neural network