Nonsmooth optimization via quasi-Newton methods

By A. S. Lewis, Michael L. Overton. Mathematical programming, 2012

396 citations recorded by the source corpus.

Subfields: Computational Theory and Mathematics, Numerical Analysis

View Paper: https://www.semanticscholar.org/paper/8aedae3fe057e4eb4387801ba2dbd0f5c6a98616

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Theory and Mathematics

In Numerical Analysis

Citations in the corpus

Batch learning from logged bandit feedback through counterfactual risk minimization (2015, background): “In particular, prior work (Yu et al., 2010; Lewis and Overton, 2013) has established theoretically sound modiﬁcations to L-BFGS for non-smooth non-convex optimization.”
The Self-Normalized Estimator for Counterfactual Learning (2015, background): “The choice of L-BFGS for non-convex and non-smooth optimization is well supported [25, 26].”
Shampoo: Preconditioned Stochastic Tensor Optimization (2018, methodology): “Classic algorithms in this family include Newton’s method, which employs the local Hessian as a preconditioner, as well as a plethora of quasi-Newton methods (e.g., [8, 15, 19]) that can be used whenever second-order information is unavailable or too expensive to compute.”
Stein Unbiased GrAdient estimator of the Risk (SUGAR) for Multiple Parameter Selection (2014, methodology): “For quasi-Newton optimization, we used the Broyden-FletcherGoldfarb-Shanno (BFGS) method with the implementation of [43].”
Elements of style (2015, background): “While the notion of style extends beyond shape, we consider a purely geometry based measure; for most modeling applications properties such as texture can be easily changed once a shape is available.”
A Nonlinear Optimization Approach to H2-Optimal Modeling and Control (2013, methodology): “However, in practice, these methods have been shown to work well on certain non-smooth problems as well, see for example Lewis and Overton [2012]. The line search strategy is to find a direction pk , and a step αk , such that fk f (xk) > f (xk + αkpk).”