A Line-Search Descent Algorithm for Strict Saddle Functions with\n Complexity Guarantees Article Swipe
Related Concepts
Backtracking
Descent (aeronautics)
Algorithm
Saddle
Line (geometry)
Property (philosophy)
Line search
Rank (graph theory)
Matrix (chemical analysis)
Matching (statistics)
Computer science
Descent direction
Gradient descent
Saddle point
Mathematics
Mathematical optimization
Combinatorics
Path (computing)
Artificial intelligence
Engineering
Epistemology
Geometry
Composite material
Materials science
Statistics
Artificial neural network
Aerospace engineering
Philosophy
Programming language
Michael O’Neill
,
Stephen J. Wright
·
YOU?
·
· 2020
· Open Access
·
· DOI: https://doi.org/10.48550/arxiv.2006.07925
· OA: W4287757759
YOU?
·
· 2020
· Open Access
·
· DOI: https://doi.org/10.48550/arxiv.2006.07925
· OA: W4287757759
We describe a line-search algorithm which achieves the best-known worst-case\ncomplexity results for problems with a certain "strict saddle" property that\nhas been observed to hold in low-rank matrix optimization problems. Our\nalgorithm is adaptive, in the sense that it makes use of backtracking line\nsearches and does not require prior knowledge of the parameters that define the\nstrict saddle property.\n
Related Topics
Finding more related topics…