Dynamical Systems
-
A complicated system is really only sensitive to a few parameters in a space whose dimension can be determined (perhaps not very reliably)
-
Parameters (assumed to be) governed by coupled differential equations ----> chaos
|
Statistical Systems
-
Don't ignore the (1023) (other) degrees of freedom but rather sum over them subject to constraints ---> maximize information, entropy ..........
|
Are approaches consistent?
-
Thermodynamics does show how macroscopic variables emerge from a statistical formulation. How ever only a qualitative relation between dynamical and statistical systems ?
-
"Back-propagation" neural networks unbiased parameterization
|
