On the Barzilai-Borwein Method
By Roger Fletcher. 2005
316 citations recorded by the source corpus.
Subfields: Computational Mechanics, Computational Theory and Mathematics, Numerical Analysis
View Paper: https://www.semanticscholar.org/paper/44a332b8dcbfd03cd704ed2dbae52ed2b5f53e20
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Computational Mechanics
- Tom Goldstein
- Yuhong Dai
- Christoph Studer
- Shiqian Ma
- Roger Fletcher
- G. Frassoldati
- Gaetano Zanghirati
- Luca Zanni
In Computational Theory and Mathematics
- Tom Goldstein
- Yuhong Dai
- William W. Hager
- Hongchao Zhang
- G. Frassoldati
- Gaetano Zanghirati
- Luca Zanni
- Christoph Studer
In Numerical Analysis
- Yuhong Dai
- William W. Hager
- Hongchao Zhang
- Klaus Schittkowski
- G. Frassoldati
- Gaetano Zanghirati
- Luca Zanni
- Shiqian Ma
Citations in the corpus
- A Field Guide to Forward-Backward Splitting with a FASTA Implementation (2014, background): “A number of variations of spectral descent are reviewed by Fletcher [38], several of which perform better in practice than the “classic” stepsize rules (32).”
- Label Propagation via Teaching-to-Learn and Learning-to-Teach (2017, methodology): “Finally, a suitable step size τ is found by Barzilai–Borwein method [37], so that the optimal S ( iter + 1 ) can be located (Lines 7–11).”
- New adaptive stepsize selections in gradient methods (2008, methodology): “To analyse the convergence of any gradient method for a quadratic function, we can assume without loss of generality that A is diagonal with distinct eigenvalues [13]: A = diag(λ1, λ2, .”
- Projected Barzilai-Borwein methods for large-scale box-constrained quadratic programming (2005, methodology): “(2005) 100: 21–47 Numerische Mathematik Projected Barzilai-Borwein methods for large-scale box-constrained quadratic programming Yu-Hong Dai1, Roger Fletcher2 1 State Key Laboratory of Scientific and Engineering Computing, Institute of Computational Mathematics and Scientific/Engineering computing, ”
- The cyclic Barzilai-–Borwein method for unconstrained optimization (2006, methodology): “The superior performance of cyclic SD, compared to the ordinary SD, as shown in Dai & Fletcher (2005a), leads us to consider the CBB method: αm +i = αBBm +1 for i = 1, . . . , m, (1.7) where m 1 is again the cycle length.”
- Barzilai-Borwein Step Size for Stochastic Gradient Descent (2016, methodology): “For convergence analysis, generalizations and variants of the BB method, we refer the interested readers to [14, 15, 6, 9, 4, 5, 3] and references therein.”
- Steepest descent (2010, background): “For an excellent overview on this subject and further details see [7].”