Fixed points of arbitrarily deep 1-dimensional neural networks.

Andrew Cook Andy Hammerlindl Warwick Tucker

Published in: CoRR (2023)

Keyphrases

fixed point
neural network
sufficient conditions
temporal difference learning
dynamical systems
linear logic
initial conditions
approximate value iteration
belief propagation
stationary points
computer vision
generative model
asymptotic properties
unsupervised learning
message passing
free energy