EURO-Online login
- New to EURO? Create an account
- I forgot my username and/or my password.
- Help with cookies
(important for IE8 users)
3450. First-order Trust-region Methods with Adaptive Sampling
Invited abstract in session TA-34: New Algorithms for Nonlinear Optimization, stream Advances in large scale nonlinear optimization.
Tuesday, 8:30-10:00Room: 43 (building: 303A)
Authors (first author is the speaker)
| 1. | Sara Shashaani
|
| Industrial and Systems Engineering, North Carolina State University |
Abstract
We are interested in optimizing noisy black-box problems that are non-convex in expectation. Such settings occur in a multitude of engineering and machine learning applications. For the noisy black-boxes that provide direct derivative observations, we devise new variants of trust-region methods that use adaptive sampling schemes and adjusted acceptance criteria to improve the iteration and sample complexity of these algorithms. The resulting methods will have clear implementable guidelines for speedy and robust finite-time performance.
Keywords
- Continuous Optimization
- Stochastic Optimization
- Efficiency Analysis
Status: accepted
Back to the list of papers