Estimate sequence methods: extensions and approximations

By Michel Baes. 2009

125 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/b1eb23cd74e3a632f9df132005f6bad6d013af69

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

Lower Bounds for Higher-Order Convex Optimization (2017, background): “Higher-Order (Baes, 2009) Yes k > 2 Bounded Lk O (( Lk+1 ε )1/(k+1))”
On Lower and Upper Bounds in Smooth and Strongly Convex Optimization (2016, background): “Eventually, it was this line of inquiry that led to the discovery of AGD by Nesterov [16], a slight modification of the standard GD algorithm, whose iteration complexity is O (√ κ ln(1/ ) ) .”
A variational perspective on accelerated methods in optimization (2016, methodology): “This auxiliary sequence is essentially that used by Nesterov in his constructions of accelerated mirror descent [26] and accelerated cubic-regularized Newton’s method [28], and later generalized by Baes [6].”
Accelerated proximal stochastic dual coordinate ascent for regularized loss minimization (2013, background): “The rest of the proofs are provided in the appendix.”
Convergence Rates of Inexact Proximal-Gradient Methods for Convex Optimization (2011, methodology): “…of a ﬁxed deterministic error in the gradient, and shown that accelerated gradient methods achieve the optimal convergence rate up to some accuracy that depends on the ﬁxed error level [20, 21, 22], while the earlier work of [23] analyzes the gradient method in the context of a ﬁxed error level.”
On the influence of momentum acceleration on online learning (2016, methodology): “al., 2013; Kahou et al., 2013; Szegedy et al., 2015; Zareba et al., 2015). The performance of Nesterov’s acceleration with deterministic and bounded gradient error was examined in (d’Aspremont, 2008; Baes, 2009; Devolder et al., 2014; Lessard et al., 2016). The source of the inaccuracy in the gradie”
A Lyapunov Analysis of Momentum Methods in Optimization (2016, methodology): “The first is the version introduced by Baes [2]: Algorithm 5 Accelerated Mirror Descent (Weakly Convex Setting) Assumptions: f ,h are convex and differentiable.”