Scalar Holder Heavy Tail Stationarity — Scalar Hölder Stochastic Stationarity
Descriptions may be stale. Consult the linked Lean sources.
Scalar Hölder Stochastic Stationarity · Lean proof
Scalar
Hölder Stochastic Stationarity
We consider a differentiable, possibly nonconvex function with a globally
Hölder continuous derivative. A given point has objective gap at
most above a global
lower bound. At , the oracle
returns an unbiased scalar estimate of using a fresh independent
sample; its error has uniformly bounded th absolute moment for , allowing heavy-tailed noise.
For , the goal is a
measurable output satisfying ,
where is the joint law of the
samples and the algorithm’s independent randomness.
The underlying optimization problem is with
accuracy measured by stationarity. Fix a measurable sample space and parameters The
algorithm receives , , , , , , , and ; the objective and sampling
distribution are accessed through the oracle.
Our assumptions:
Assumption 1 (Hölder continuity). The function is differentiable everywhere, and
Assumption 2 (Lower bound and initial gap). There
exists such that
Assumption 3 (Independent stochastic derivative
oracle). There exist a probability measure on and a jointly measurable function
.
For each , both and are integrable
under , and At call , a query receives the scalar reply
. The samples are measurable, mutually
independent, and each has law .
The algorithm’s randomness is independent of the entire sample
sequence.
Queries may be made at any point of . After the allotted oracle
calls, the output must be
measurable and satisfy
Proved result
OracleTemplate.Convergence LongTails.Setups.ScalarHolderStochastic.spec OracleTemplate.ScalarHolderStochastic.budget
Relevant previous results
No qualifying prior reference verified; this is not a proof of novelty.