Zero-one law for directional transience in dimension at least three
OpenOAI.DirectionalTransience.uniformElliptic_directional_zero_oneopenai-comparatorprobabilityrandom-walk-in-random-environment
Let , let be a uniformly elliptic probability law on transition rows, and let . Then the annealed probability that the i.i.d. random walk in random environment is transient in direction (i.e. ) is either or .
Preamble
import Mathlib import Definitions.Def_VelocityHemisphere namespace OAI open MeasureTheory ProbabilityTheory Filter open scoped ENNReal NNReal BigOperators Topology namespace DirectionalTransience
Formal statement
theorem uniformElliptic_directional_zero_one {d : ℕ} (hd : 3 ≤ d)
(ν : Measure (Row d)) [IsProbabilityMeasure ν] (hue : UniformElliptic ν)
(ℓ : Vector d) (hℓ : ℓ ≠ 0) :
annealedLaw ν (TransientPaths ℓ) = 0 ∨ annealedLaw ν (TransientPaths ℓ) = 1 := by
sorry
end DirectionalTransience
end OAI
Source
OpenAI, openai/math commit adc7f1241b42e322a6451854ab7e4b4c146bf78a, lean/OAI/Probability/Ballisticity/PositiveEscape.lean, lemma `uniformElliptic_directional_zero_one` (line 10), obtained there from the directional zero-one law lean/OAI/Probability/DirectionalWalk (platform theorem `OAI.DirectionalZeroOne.directional_zero_one`) via the row bridge lean/OAI/Probability/Ballisticity/RowBridge.lean. It is used for the velocity hemisphere theorem (lean/OAI/Probability/Ballisticity/VelocityHemisphere.lean) of the manuscript https://github.com/openai/math/blob/adc7f1241b42e322a6451854ab7e4b4c146bf78a/preprints/Directional-transience-implies-ballisticity-September-23-2026/paper.pdf.