Giter Club home page Giter Club logo

manifolds.jl's Introduction

Manifolds

Define an interface for working on manifolds, as well as a set of predefined manifolds. The target use is optimization of differential equation integration that stay on a specific manifold at each step of the algorithm.

Example

using Manifolds
M = Sphere()
x = randn(3)
@assert norm(retract(M, x)) ≈ 1

Interface

The following functions are supported:

  • retract!(M::Manifold, x): retract x, i.e. find a point close to x that belongs to the specified manifold
  • project_tangent!(M::Manifold, g, x): project g onto the tangent space at x as well as out-of-place versions retract(M,x) and project_tangent(M, g, x).

Supported manifolds

The following manifolds are currently supported:

  • Flat: Euclidean space, default. The operations are no-ops.
  • Sphere: spherical constraint ||x|| = 1.
  • Stiefel: Stiefel manifold of N by n matrices with orthogonal columns, i.e. X'*X = I. The default algorithm for retractions on the manifold is to use a SVD, but a faster (although less stable) algorithm is available by calling Stiefel(:CholQR).

The following meta-manifolds construct manifolds out of pre-existing ones:

  • PowerManifold: identical copies of a specified manifold
  • ProductManifold: product of two (potentially different) manifolds

Implementing new manifolds is as simple as adding methods project_tangent!(M::YourManifold,g,x) and retract!(M::YourManifold,x). If you implement another manifold, please contribute a PR!

References

The Geometry of Algorithms with Orthogonality Constraints, Alan Edelman, Tomás A. Arias, Steven T. Smith, SIAM. J. Matrix Anal. & Appl., 20(2), 303–353

Optimization Algorithms on Matrix Manifolds, P.-A. Absil, R. Mahony, R. Sepulchre, Princeton University Press, 2008

manifolds.jl's People

Contributors

antoine-levitt avatar

Stargazers

Kamil Ziemian avatar

Watchers

 avatar James Cloos avatar

manifolds.jl's Issues

Implicit Manifolds

In DiffEq we define implicit manifolds and project using a Newton method here:

https://github.com/JuliaDiffEq/DiffEqCallbacks.jl/blob/master/src/manifold.jl

We set this up with the generic nlsolve interface:

http://docs.juliadiffeq.org/latest/features/linear_nonlinear.html#Nonlinear-Solvers:-nlsolve-Specification-1

which defaults to using NLsolve.jl, this extra part can easily be dropped. Basically it's just that g is a residual function and you project to this manifold by calling a non-linear solver, not much more to it.

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.