Stochastic Search In Changing Situations

• Type: Article in International Conference Proceedings Book
• Date: 2017

Authors

Abdolmaleki, A (Author) Other; Simães, DA (Author) Other; lau, n (Author) FCUP; reis, lp (Author) REIT; Price, B (Author) Other; Neumann, G (Author) Other

Conference proceedings International

Title: The Workshops of the The Thirty-First AAAI Conference on Artificial Intelligence, Saturday, February 4-9, 2017, San Francisco, California, USA

Pages: 636-642

31st AAAI Conference on Artificial Intelligence, AAAI 2017

4 February 2017 through 5 February 2017

Indexing

Scopus - 0 Citations

Other information

Authenticus ID: P-00N-YJ4

Abstract (EN): Stochastic search algorithms are black-box optimizer of an objective function. They have recently gained a lot of attention in operations research, machine learning and policy search of robot motor skills due to their ease of use and their generality. However, when the task or objective function slightly changes, many stochastic search algorithms require complete re-leaming in order to adapt thesolution to the new objective function or the new context. As such, we consider the contextual stochastic search paradigm. Here, we want to find good parameter vectors for multiple related tasks, where each task is described by a continuous context vector. Hence, the objective function might change slightly for each parameter vector evaluation. In this paper, we investigate a contextual stochastic search algorithm known as Contextual Relative Entropy Policy Search (CREPS), an information-theoretic algorithm that can learn from multiple tasks simultaneously. We show the application of CREPS for simulated robotic tasks.

Language: English

Type (Professor's evaluation): Scientific