CS Colloquium Series @ UCY

Department of Computer Science - University of Cyprus

The Department of Computer Science at the University of Cyprus holds research colloquiums and social hours approximately once weekly. All university students, faculty, and staff are invited to attend. Notifications about new and upcoming events are automatically disseminated to a variety of institutional lists.
If you don't receive these notifications, but want to get informed about upcoming colloquium announcements, you can do the following:
mail List rss RSS Directions Directions

Colloquium Coordinator: Demetris Zeinalipour

Colloquium: Boosting for Probability Estimation & Cost-Sensitive Learning, Dr. Nikolaos Nikolaou (University of Manchester, UK), Thursday, November 16, 2017, 15:00-16:00 EET.


The Department of Computer Science at the University of Cyprus cordially invites you to the Colloquium entitled:

Boosting for Probability Estimation & Cost-Sensitive Learning

 

Speaker: Dr. Nikolaos Nikolaou
Affiliation: University of Manchester, UK
Category: Colloquium
Location: Room 148, Faculty of Pure and Applied Sciences (FST-01), 1 University Avenue, 2109 Nicosia, Cyprus (directions)
Date: Thursday, November 16, 2017
Time: 15:00-16:00 EET
Host: Chris Christodoulou (cchrist-AT-cs.ucy.ac.cy)
URL: https://www.cs.ucy.ac.cy/colloquium/index.php?speaker=cs.ucy.2017.nikolaou

Abstract:
We provide a unifying perspective for two decades of work on cost-sensitive Boosting algorithms. We critique the relevant literature -consisting of more than 15 variants of the original algorithm- using four theoretical frameworks: Bayesian decision theory, functional gradient descent, margin theory, and probabilistic modelling. We find that only 3 of the published Adaboost variants are consistent with the rules of all the frameworks —and even they require their outputs to be calibrated to achieve this. Experiments on 18 datasets across 21 degrees of imbalance support the hypothesis -showing that once calibrated, they perform equivalently, and outperform all others. Our final recommendation -based on simplicity, flexibility and performance- is to use the original Adaboost algorithm with a shifted decision threshold and calibrated probability estimates. We then move on to the online setting which imposes the additional complication of having to decide whether to use new datapoints to update the parameters of the ensemble or those of the calibrator function. We propose resolving this decision with the aid of bandit optimization algorithms and present initial results suggesting superior performance to uncalibrated and naively-calibrated online boosting ensembles.

Short Bio:
Dr. Nikolaos Nikolaou is an EPSRC Doctoral Prize Fellow at the School of Computer Science of the University of Manchester. He received his Electronic & Computer Engineering Diploma from the Technical University of Crete in 2011 and his M.Sc. in Artificial Intelligence from the University of Edinburgh in 2012, supported by a PSAS Award Scholarship. He received his Ph.D. from the University of Manchester in 2016, funded by the EPSRC. His doctoral thesis was in the area of cost-sensitive boosting algorithms but his research interests included multi-label classification & feature selection. His current research focuses on ensemble learning, online learning & information-theory, with applications to renewable energy and memory controller design.

  Web: https://www.cs.ucy.ac.cy/colloquium/
  Mailing List: https://listserv.cs.ucy.ac.cy/mailman/listinfo/cs-colloquium
  RSS: https://www.cs.ucy.ac.cy/colloquium/rss.xml
  Calendar: https://www.cs.ucy.ac.cy/colloquium/schedule/cs.ucy.2017.nikolaou.ics