Deriving and improving CMA-ES with information geometric trust regions

CMA-ES is one of the most popular stochastic search algorithms. It performs favourably in many tasks without the need of extensive parameter tuning. The algorithm has many beneficial properties, including automatic step-size adaptation, efficient covariance updates that incorporates the current samp...

Full description

Bibliographic Details
Main Author: Abdolmaleki, Abbas (author)
Other Authors: Price, Bob (author), Lau, Nuno (author), Reis, L. P. (author), Neumann, Gerhard (author)
Format: conferencePaper
Language:eng
Published: 2017
Subjects:
Expectation maximisation
Stochastic search
Trust regions
Ciências Naturais::Ciências da Computação e da Informação
Science & Technology
Online Access:http://hdl.handle.net/1822/51451
Country:Portugal
Oai:oai:repositorium.sdum.uminho.pt:1822/51451

Internet

http://hdl.handle.net/1822/51451

Similar Items