Module high_dimensional 

High-dimensional optimization methods

This module provides optimization algorithms specifically designed for high-dimensional problems where standard methods may be too expensive.

Modules

• coordinate_descent: Coordinate descent variants (cyclic, randomized, greedy, proximal, block)
• kaczmarz: Kaczmarz iteration for solving linear systems Ax = b
• sketched_gd: Sketched gradient descent with dimensionality reduction

Re-exports

pub use coordinate_descent::BlockCoordinateDescent;

pub use coordinate_descent::CoordinateDescentConfig;

pub use coordinate_descent::CoordinateDescentResult;

pub use coordinate_descent::CoordinateDescentSolver;

pub use coordinate_descent::CoordinateSelectionStrategy;

pub use coordinate_descent::ProximalCoordinateDescent;

pub use coordinate_descent::RegularizationType;

pub use kaczmarz::BlockKaczmarz;

pub use kaczmarz::ExtendedKaczmarz;

pub use kaczmarz::KaczmarzConfig;

pub use kaczmarz::KaczmarzResult;

pub use kaczmarz::KaczmarzSolver;

pub use kaczmarz::KaczmarzVariant;

pub use sketched_gd::SketchType;

pub use sketched_gd::SketchedGdConfig;

pub use sketched_gd::SketchedGdResult;

pub use sketched_gd::SketchedGradientDescent;

Modules

coordinate_descent

Coordinate Descent optimization methods

kaczmarz

Kaczmarz method for solving linear systems Ax = b

sketched_gd

Sketched Gradient Descent