Symmetric Reverse Fisher Global GD — Gradient Descent for Reverse Fisher Estimation
Descriptions may be stale. Consult the linked Lean sources.
Gradient Descent for Reverse Fisher Estimation · Lean proof
Gradient Descent for
Reverse Fisher Estimation
We consider estimating a hidden vector , where , from exact gradients of the
reverse Fisher objective : the
mean squared difference between candidate and target mixture scores,
averaged under the candidate symmetric Gaussian mixture with identity
covariance. The signal satisfies , where . The oracle returns at . We analyze gradient descent with
positive steps and algorithm-chosen initialization. For , the goal is a
measurable output
satisfying
with probability at least
over the algorithm’s randomness.
Equip with its
Euclidean inner product and norm. For a candidate parameter , define the
symmetric mixture and its score by where is
the identity matrix. The objective is The associated optimization problem is The algorithm receives ; the target
is hidden.
Our assumptions:
Assumption 1 (Signal bounds). The fixed target
belongs to the prescribed Euclidean annulus:
Our result is algorithm-specific: we analyze gradient descent. The
algorithm chooses its initialization , possibly at
random. At each iteration it queries its current point and updates
according to The step may depend on the algorithm’s
randomness and information available before the current oracle reply.
Queries may be made anywhere in . After oracle calls, the output is the
current iterate .
The required accuracy concerns recovery up to sign. The output must
be measurable and satisfy where is the law of the algorithm’s
randomness.
Proved result
OracleTemplate.Convergence LongTails.Setups.SymmetricGaussianReverseFisher.spec
LongTails.Theorems.SymmetricReverseFisherGlobalGD.budget
Relevant previous results
No qualifying prior reference verified; this is not a proof of novelty.