Expressive power of first-order recurrent neural networks determined by their attractor dynamics.
Jérémie CabessaAlessandro E. P. VillaPublished in: J. Comput. Syst. Sci. (2016)
Keyphrases
- expressive power
- recurrent neural networks
- first order logic
- recurrent networks
- dynamical systems
- neural network
- data complexity
- relational algebra
- dynamic behavior
- transitive closure
- feed forward
- computational properties
- cellular automata
- reservoir computing
- query language
- nonlinear dynamic systems
- knowledge representation
- propositional logic
- artificial neural networks
- fixed point
- echo state networks
- relational calculus
- closure properties
- theorem prover
- aggregate functions
- database
- knowledge base
- recursive queries
- quantifier elimination
- data model
- relation algebra
- databases