N
Nikos Karampatziakis
Researcher at Microsoft
Publications - 52
Citations - 2577
Nikos Karampatziakis is an academic researcher from Microsoft. The author has contributed to research in topics: Multiclass classification & Randomized algorithm. The author has an hindex of 19, co-authored 50 publications receiving 2332 citations. Previous affiliations of Nikos Karampatziakis include Cornell University & University of Houston.
Papers
More filters
Proceedings ArticleDOI
An empirical evaluation of supervised learning in high dimensions
TL;DR: To the surprise, the method that performs consistently well across all dimensions is random forests, followed by neural nets, boosted trees, and SVMs, and the effect of increasing dimensionality on the performance of the learning algorithms changes.
Journal ArticleDOI
Three-Dimensional Face Recognition in the Presence of Facial Expressions: An Annotated Deformable Model Approach
Ioannis A. Kakadiaris,G. Passalis,George Toderici,Mohammed N. Murtuza,Yunliang Lu,Nikos Karampatziakis,Theoharis Theoharis +6 more
TL;DR: This paper presents the computational tools and a hardware prototype for 3D face recognition and presents the results on the largest known, and now publicly available, face recognition grand challenge 3D facial database consisting of several thousand scans.
Proceedings Article
Gradient Coding: Avoiding Stragglers in Distributed Learning
TL;DR: This work proposes a novel coding theoretic framework for mitigating stragglers in distributed learning and shows how carefully replicating data blocks and coding across gradients can provide tolerance to failures andstragglers for synchronous Gradient Descent.
Proceedings Article
Efficient optimal learning for contextual bandits
Miroslav Dudík,Daniel Hsu,Satyen Kale,Nikos Karampatziakis,John Langford,Lev Reyzin,Tong Zhang +6 more
TL;DR: This work provides the first efficient algorithm with an optimal regret and uses a cost sensitive classification learner as an oracle and has a running time polylog(N), where N is the number of classification rules among which the oracle might choose.
Posted Content
Efficient Optimal Learning for Contextual Bandits
Miroslav Dudík,Daniel Hsu,Satyen Kale,Nikos Karampatziakis,John Langford,Lev Reyzin,Tong Zhang +6 more
TL;DR: In this article, a cost sensitive classification learner is used as an oracle and an algorithm with an optimal regret is presented. But regret is not optimal in the online setting, and the regret is additive rather than multiplicative in feedback delay.