Skip navigation
Skip navigation

Defensive universal learning with experts

Poland, Jan; Hutter, Marcus

Description

This paper shows how universal learning can be achieved with expert advice. To this aim, we specify an experts algorithm with the following characteristics: (a) it uses only feedback from the actions actually chosen (bandit setup), (b) it can be applied with countably infinite expert classes, and (c) it copes with losses that may grow in time appropriately slowly. We prove loss bounds against an adaptive adversary. From this, we obtain a master algorithm for “reactive” experts problems, which...[Show more]

CollectionsANU Research Publications
Date published: 2005
Type: Conference paper
URI: http://hdl.handle.net/1885/15047
DOI: 10.1007/11564089_28

Download

File Description SizeFormat Image
Poland and Hutter Defensive Universal Learning 2005.pdf249.15 kBAdobe PDFThumbnail


Items in Open Research are protected by copyright, with all rights reserved, unless otherwise indicated.

Updated:  12 November 2018/ Responsible Officer:  University Librarian/ Page Contact:  Library Systems & Web Coordinator