Mathematical Study of First-order Algorithms for Big Datasets
Grant Data
Project Title
Mathematical Study of First-order Algorithms for Big Datasets
Principal Investigator
Professor Yuan, Xiaoming
(Principal Investigator (PI))
Duration
36
Start Date
2018-11-01
Completion Date
2021-10-31
Amount
684585
Conference Title
Mathematical Study of First-order Algorithms for Big Datasets
Keywords
Big Dataset, First-order Algorithms, Numerical Algorithms, Numerical Optimization
Discipline
Applied Mathematics
Panel
Physical Sciences (P)
HKU Project Code
12302318
Grant Type
General Research Fund (GRF)
Funding Year
2018
Status
Completed
Objectives
1 To design a series of new algorithms, based on some benchmark first-order methods, for big datasets of some important optimization models. 2 To rigorously prove the convergence and estimate the sublinear convergence rates for the new algorithms. 3 To introduce some novel theoretical and analytical frameworks for studying the linear convergence rates of the new algorithms. 4 To apply the new algorithms to some important and popular big datasets and to tackle crucial implementation issues.