IEEE transactions on information theory, 2012-07, Vol.58 (7), p.4164-4175
2012
Volltextzugriff (PDF)
Details
- Autor(en) / Beteiligte
- Titel
- Interior-Point Methods for Full-Information and Bandit Online Learning
- Ist Teil von
- IEEE transactions on information theory, 2012-07, Vol.58 (7), p.4164-4175
- Ort / Verlag
- New York, NY: IEEE
- Erscheinungsjahr
- 2012
- Quelle
- IEEE Electronic Library (IEL)
- Beschreibungen/Notizen
- We study the problem of predicting individual sequences with linear loss with full and partial (or bandit) feed- back. Our main contribution is the first efficient algorithm for the problem of online linear optimization in the bandit setting which achieves the optimal Õ(√(T)) regret. In addition, for the full-information setting, we give a novel regret minimization algorithm. These results are made possible by the introduction of interior-point methods for convex optimization to online learning.
- Sprache
- Englisch
- Identifikatoren
- ISSN: 0018-9448eISSN: 1557-9654DOI: 10.1109/TIT.2012.2192096Titel-ID: cdi_proquest_journals_1021021234
- Format
- –
- Schlagworte
- Applied sciences, Bandit feedback, Coding theory, Convex functions, Decision making, Ellipsoids, Exact sciences and technology, Feedback, Information theory, Information, signal and communications theory, interior-point methods, Linear codes, Minimization, Newton method, online convex optimization, Online instruction, online learning, Optimization, Optimization algorithms, Telecommunications and information theory, Vectors