Publications

Citation

BibTex format

@inproceedings{Gyorgy:2016,
author = {Gyorgy, A and Szepesvari, C},
pages = {2943--2951},
publisher = {Journal of Machine Learning Research},
title = {Shifting Regret, Mirror Descent, and Matrices},
url = {http://hdl.handle.net/10044/1/31670},
year = {2016}
}
Download

RIS format (EndNote, RefMan)

TY  - CPAPER
AB  - We consider the problem of online prediction inchanging environments. In this framework theperformance of a predictor is evaluated as theloss relative to an arbitrarily changing predictor,whose individual components come from a baseclass of predictors. Typical results in the literatureconsider different base classes (experts, linearpredictors on the simplex, etc.) separately.Introducing an arbitrary mapping inside the mirrordecent algorithm, we provide a frameworkthat unifies and extends existing results. As anexample, we prove new shifting regret bounds formatrix prediction problems.
AU  - Gyorgy,A
AU  - Szepesvari,C
EP  - 2951
PB  - Journal of Machine Learning Research
PY  - 2016///
SN  - 1532-4435
SP  - 2943
TI  - Shifting Regret, Mirror Descent, and Matrices
UR  - http://hdl.handle.net/10044/1/31670
ER  -
Download