Generalized Smooth Convex Mirror Geometry
Descriptions may be stale. Consult the linked Lean sources.
Generalized Smooth Convex Mirror Geometry
We minimize a differentiable convex function
The optimization problem is
Our assumptions:
Assumption 1 (Convexity and generalized smoothness).
The function
Assumption 2 (Mirror geometry and proximal access).
The function
Assumption 3 (Initial bounds). There exists
The oracle returns
After the allotted calls, the output
Results
Convex differentiable
| Result | Oracle complexity | Novelty | Limitations |
|---|---|---|---|
| (1) DynamicMirrorGeneralizedAcceleration
(Ours) |
|||
| (2) Accelerated mirror descent (Yu et al. 2026), Thm. 3.7 | (1): |
(1): affine links and unconstrained domains only; three calls per iteration instead of two. | |
| (3) AGD (Tyurin 2026), Thm. 4.3; Euclidean quadratic mirror | (1): scope extends Euclidean geometry to arbitrary norms with strongly convex mirrors; matches the accelerated accuracy exponent. | (1): lacks logarithm
saturation below |
|
| (4) (Rudomanov et al. 2026), Thm. 3.7; Euclidean quadratic mirror | (1): scope extends Euclidean geometry to arbitrary norms with strongly convex mirrors. | (1): needs |
Abbr: