Minimization Methods For Non Differentiable Functions

By Franziska Wulf. 2016

500 citations recorded by the source corpus.

How researchers see this paper

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

Citations in the corpus

The Value of Collaboration in Convex Machine Learning with Differential Privacy (2019, background): “The existence of sub-differentials is guaranteed for convex functions [30].”
Convergence Rates for Deterministic and Stochastic Subgradient Methods Without Lipschitz Continuity (2017, background): “Shor used the same elementary argument underlying Theorem 5 to show for any minimizer x∗ ∈ X∗: either some k = 0 . . .”
Stochastic Approximations and Perturbations in Forward-Backward Splitting for Monotone Operators (2015, methodology): “Stochastic approximations in the gradient method were then investigated in the Russian literature of the late 1960s and early 1970s [27, 28, 29, 33, 42, 49].”
Boosted Bellman Residual Minimization Handling Expert Demonstrations (2014, background): “As presented by [10], a boosting algorithm is a projected sub-gradient descent [22] of a convex functional in a specific functions space (here RS×A) which has to be a Hilbert space.”
Structured Pruning for Deep Convolutional Neural Networks: A Survey (2023, methodology): “To optimize the non-smooth `1 penalty term, a subgradient descent [228] is used.”
Phase Transitions in Transfer Learning for High-Dimensional Perceptrons (2021, unknown): “Following [33], the sub-differential set ∂L(·) can be expressed as follows:”
A Differential Equation for Modeling Nesterov's Accelerated Gradient Method: Theory and Insights (2014, background): “First-order methods have regained popularity as data sets and problems are ever increasing in size and, consequently, there has been much research on the theory and practice of accelerated ﬁrst-order schemes.”
Parallel Direction Method of Multipliers (2014, unknown): “If ρ = 0, the primal update can be solved in a parallel block coordinate fashion [3, 19], leading to the dual ascent method.”
Boosted and reward-regularized classification for apprenticeship learning (2014, background): “As presented in [7], a boosting algorithm is a projected sub-gradient descent [20] of a convex functional in a specific functions space (here RS×A) which has to be a Hilbert space.”