Power Smooth Anchoring — Stochastic Power-Smooth Fixed Points
Descriptions may be stale. Consult the linked Lean sources.
Stochastic Power-Smooth Fixed Points · Lean proof
Stochastic Power-Smooth Fixed
Points
We seek a fixed point of a nonexpansive map on a separable real Banach space
whose norm is power-smooth with
exponent and constant
. A given point lies within distance of a fixed point . At any query , the oracle returns an unbiased
estimate of using an
independent sample, with error th
moment at most , where
; this permits
heavy-tailed noise. For , the goal is a
measurable output with
integrable residual satisfying ,
where is the joint law of
algorithm and oracle randomness.
The fixed-point problem is to find such that Equip
with its Borel sigma-algebra, and
let be the measurable sample
space. The algorithm receives ,
, , , , , , and ; it accesses through the stochastic oracle described
below.
Our assumptions:
Assumption 1 (Power-smooth norm). There exists a map
, where is the continuous real dual of , such that
Assumption 2 (Nonexpansiveness and initial
distance). The map satisfies
There exists such that
Assumption 3 (Unbiased oracle with bounded moment).
Let be a probability measure on
, and let be jointly measurable.
For every , the random vector
is Bochner integrable
under , the scalar is integrable, and
At call , a query receives . The measurable samples are mutually independent
with common law , and the entire
sample sequence is independent of the algorithm’s randomness.
Queries may be made anywhere in . After the allotted oracle calls, the
output must be measurable,
its residual
must be integrable under , and the
accuracy requirement is
Proved result
OracleTemplate.Convergence LongTails.Setups.StochasticPowerSmoothFixedPoint.spec
LongTails.Theorems.PowerSmoothAnchoring.budget
Relevant previous results
No qualifying prior reference verified; this is not a proof of novelty.