First-Order Algorithms Converge Faster than $O(1/k)$ on Convex Problems.
Ching-pei LeeStephen J. WrightPublished in: ICML (2019)
Keyphrases
- benchmark problems
- optimization problems
- convex optimization problems
- test problems
- computational problems
- related problems
- computational complexity
- significant improvement
- search methods
- convex programs
- convex programming
- problems in computer vision
- convex functions
- approximate solutions
- difficult problems
- convex optimization
- combinatorial optimization
- higher order
- computational cost
- search space
- data mining
- semi definite programming
- minimization problems
- learning algorithm
- computationally demanding
- objective function
- search algorithm
- specific problems
- memory efficient
- theoretical analysis
- practical problems
- orders of magnitude
- globally optimal
- highly efficient