Learning Faster from Easy Data
In online learning algorithms are typically analysed in the worst case. However, not all data are created equal. We aim to quantify the difficulty of the data, and design efficient algorithms that adapt to it automatically.
-
Metagrad: Adaptation using multiple learning rates in online learning
Tim van Erven, Wouter M. Koolen, Dirk van der Hoeven
Journal of Machine Learning Research, 22(161):1–61, July 2021.
-
Lipschitz adaptivity with multiple learning rates in online learning
Zakaria Mhammedi, Wouter M. Koolen, Tim van Erven
In Proceedings of the 32nd Annual Conference on Learning Theory (COLT), pages 2490–2511, June 2019.
-
Combining adversarial guarantees and stochastic fast rates in online learning
Wouter M. Koolen, Peter Grünwald, Tim van Erven
In Advances in Neural Information Processing Systems (NeurIPS) 29, pages 4457–4465, December 2016.
-
MetaGrad: Multiple learning rates in online learning
Tim van Erven, Wouter M. Koolen
In Advances in Neural Information Processing Systems (NeurIPS) 29, pages 3666–3674, December 2016.
-
Second-order quantile methods for experts and combinatorial games
Wouter M. Koolen, Tim van Erven
In Proceedings of the 28th Annual Conference on Learning Theory (COLT), pages 1155–1175, June 2015.
-
Learning the learning rate for prediction with expert advice
Wouter M. Koolen, Tim van Erven, Peter Grünwald
In Advances in Neural Information Processing Systems (NeurIPS) 27, pages 2294–2302, December 2014.
-
Follow the leader if you can, Hedge if you must
Steven de Rooij, Tim van Erven, Peter Grünwald, Wouter M. Koolen
Journal of Machine Learning Research, 15:1281–1316, April 2014.
-
Adaptive Hedge
Tim van Erven, Steven de Rooij, Wouter M. Koolen, Peter Grünwald
In Advances in Neural Information Processing Systems (NeurIPS) 24, pages 1656–1664, December 2011.