Comparison of Echo State Networks with Simple Recurrent Networks and Variable-Length Markov Models on Symbolic Sequences
Comparison of Echo State Networks with Simple Recurrent Networks and Variable-Length Markov Models on Symbolic Sequences
M. Cernanský,P. Tiňo
2007 · DOI: 10.1007/978-3-540-74690-4_63
International Conference on Artificial Neural Networks · 16 Citations
TLDR
This work experimentally support the claim that the state space of ESN is organized according to the Markovian architectural bias principles when processing symbolic sequences.