RTR practical aspects + tCG

About
Privacy
Disclaimer

Graph Chatbot

In course

Quis elit in nisi occaecat. Id pariatur est occaecat nostrud laboris est velit. Ea sunt consequat nulla occaecat Lorem est pariatur tempor dolore fugiat ad. Deserunt in amet aliqua laboris Lorem culpa. Reprehenderit aliquip fugiat aliqua excepteur veniam veniam adipisicing consequat et veniam sunt consectetur non. Non do anim qui nisi ipsum nisi sit culpa. Mollit ex esse irure magna consectetur quis ea anim.

Description

This lecture covers practical aspects of Riemannian trust-region (RTR) optimization, including the choice of the trust-region radius and the subproblem solver, along with the truncated conjugate gradient (tCG) method for solving optimization problems. The instructor explains the key concepts, algorithms, and conditions for successful optimization using RTR and tCG.

Instructor

amet eiusmod

Fugiat exercitation consequat consequat laborum magna ex ex mollit adipisicing. Veniam sunt excepteur non fugiat incididunt officia adipisicing veniam cupidatat tempor cupidatat irure exercitation. Pariatur deserunt in incididunt est enim consequat.

Official source

Related lectures (30)

https://mediaspace.epfl.ch/media/0_snrygjwr

About this result

This page is automatically generated and may contain information that is not correct, complete, up-to-date, or relevant to your search query. The same applies to every other page on this website. Please make sure to verify the information with EPFL's official sources.

In course

DEMO: nulla deserunt nisi amet

Instructor

amet eiusmod

Related lectures (30)

Trust Region Methods: Why, with an Example

Introduces trust region methods and presents an example of Max-Cut Burer-Monteiro rank 2 optimization.

Truncated Conjugate Gradients for Trust-Region Subproblem

Explores truncated conjugate gradients for solving the trust-region subproblem in optimization on manifolds efficiently.

Optimization on Manifolds: Context and Applications

Introduces optimization on manifolds, covering classical and modern techniques in the field.

Linear convergence with Polyak-Łojasiewicz: Mechanical proof

Explores linear convergence with the Polyak-Łojasiewicz condition on a Riemannian manifold.

Newton's method on Riemannian manifolds

Covers Newton's method on Riemannian manifolds, focusing on second-order optimality conditions and quadratic convergence.