Fixed points of monotonic and (weakly) scalable neural networks.

Tomasz Piotrowski Renato L. G. Cavalcante

Published in: CoRR (2021)

Keyphrases

fixed point
neural network
approximate value iteration
sufficient conditions
dynamical systems
belief propagation
asymptotic properties
linear logic
initial conditions
competitive learning
temporal difference learning
three dimensional
lower bound
graphical models
free energy