Deterministic homogenization for fast-slow systems with chaotic noise

Consider a fast-slow system of ordinary differential equations of the form $\dot x=a(x,y)+\varepsilon^{-1}b(x,y)$, $\dot y=\varepsilon^{-2}g(y)$, where it is assumed that $b$ averages to zero under the fast flow generated by $g$. We give conditions under which solutions $x$ to the slow equations converge weakly to an It\^o diffusion $X$ as $\varepsilon\to0$. The drift and diffusion coefficients of the limiting stochastic differential equation satisfied by $X$ are given explicitly. Our theory applies when the fast flow is Anosov or Axiom A, as well as to a large class of nonuniformly hyperbolic fast flows (including the one defined by the well-known Lorenz equations), and our main results do not require any mixing assumptions on the fast flow.