OpenAI reports a directional zero–one law for finite-range environments

For every stationary ergodic, uniformly elliptic, finite-range-dependent nearest-neighbor environment on Z^d with d≥3, the annealed probability of escape in any fixed nonzero direction is either zero or one.

Curated contributions 1

Understand the problem ↓

Understanding the problem

Read the background

For every stationary ergodic, uniformly elliptic, finite-range-dependent nearest-neighbor environment on Z^d with d≥3, the annealed probability of escape in any fixed nonzero direction is either zero or one.

Why it matters

If correct, extends directional zero–one behavior beyond independent random environments.

What counts as progress

Does directional escape retain a zero–one law in finite-range-dependent random environments?

Website timeline

Publication dates and editorial changes.