2765. Adapting the coordinate-descent method for probability maximization problems
Invited abstract in session WC-31: Solution Algorithms for Optimization under Uncertainty 2, stream Stochastic and Robust optimization.
Wednesday, 12:30-14:00Room: Maurice Keyworth 1.06
Authors (first author is the speaker)
| 1. | Edit Csizmás
|
| Dept. of Informatics, John von Neumann University | |
| 2. | Rajmund Drenyovszki
|
| Dept. of Informatics, John von Neumann University | |
| 3. | Csaba Fabian
|
| Dept. of Informatics, John von Neumann University |
Abstract
For probabilistic problems, computing or estimating gradients requires a considerable effort. We build an approximate model, using a random coordinate-descent method to find new test points. We show that better results are obtained when we compute a Lipschitz constant for a region near the optimal solution.
Keywords
- Stochastic Optimization
- Convex Optimization
Status: accepted
Back to the list of papers