Evgeni Tsivtsivadze
I am postdoc at the Institute
for Computing and Information Sciences (ICIS)
Machine Learning Group, Faculty of Science
Radboud University Nijmegen
E-mail: evgeni[at]science.ru.nl
I did my PhD at the Turku Centre for Computer Science (TUCS)
Department of Information
Technology
University of Turku
My research interests are in machine learning. I am working on kernel-based methods for preference learning (ranking) and regression tasks. The application domains are bioinformatics, natural language processing, and information retrieval. I am also interested in multi-task and Bayesian learning.
Publications (bibtex file)
Kernel principal component ranking: Robust ranking on noisy data.
In Eyke Hüllermeier and Johannes Fürnkranz, editors, ECML/PKDD-Workshop on Preference Learning (PL-09), 2009.
- A Matlab implementation of the probabilistic preference learning models described in the manuscript. Download. Please cite the paper in case you are using this code. Written by Botond Cseke.
Identification of presynaptic gene clusters in synaptic signaling using functional data from genetic perturbation studies in hippocampal autapses.
In 2nd INCF Congress of Neuroinformatics, Frontiers in Neuroinformatics, 2009.
Co-regularized least-squares for label ranking.
In Eyke Hüllermeier and Johannes Fürnkranz, editors, Chapter in Preference Learning Book, 2009.
To appear.
An efficient algorithm for learning to rank from preference graphs.
Machine Learning, 75(1):129-165, 2009.
From ranking to intransitive preference learning: Rock-paper-scissors and beyond.
In Eyke Hüllermeier and Johannes Fürnkranz, editors, ECML/PKDD-Workshop on Preference Learning (PL-09), 2009.
Leanring preferences with co-regularized least squares.
In Eyke Hüllermeier and Johannes Fürnkranz, editors, ECML/PKDD-Workshop on Preference Learning (PL-08), pages 55-62, 2008.
A sparse regularized least-squares preference learning algorithm.
In Anders Holst, Per Kreuger, and Peter Funk, editors, Proceedings of the 10th Scandinavian Conference on Artificial Intelligence (SCAI 2008), volume 173, pages 76-83. IOS Press, 2008.
Regularized least-squares for learning non-transitive preferences between strategies.
In Tapani Raiko, Pentti Haikonen, and Jaakko Vyrynen, editors, Proceedings of the 13th Finnish Artificial Intelligence Conference (STeP 2008), pages 94-98, Espoo, August 2008. IOS Press.
Learning to rank with pairwise regularized least-squares.
In Thorsten Joachims, Hang Li, Tie-Yan Liu, and ChengXiang Zhai, editors, Proceedings of the SIGIR 2007 Workshop on Learning to Rank for Information Retrieval, pages 27-33, 2007.
Locality kernels for protein classification.
In Raffaele Giancarlo and Sridhar Hannenhalli, editors, Proceedings of the 7th International Workshop on Algorithms in Bioinformatics, (WABI 2007), pages 2-11. Springer, 2007.
Locality-convolution kernel and its application to dependency parse ranking.
In Moonis Ali and Richard Dapoigny, editors, Proceedings of the IEA/AIE 2006, Annecy, France, volume 4031 of Lecture Notes in Computer Science, pages 610-618. Springer, 2006.
Graph kernels versus graph representations: a case study in parse ranking.
In Thomas Gärtner, Gemma C. Garriga, and Thorsten Meinl, editors, Proceedings of the ECML/PKDD'06 workshop on Mining and Learning with Graphs (MLG 2006), Berlin, Germany, 2006.
Kernels for text analysis.
Chapter in Advances of Computational Intelligence in Industrial Systems, 116:81-97, May 2008.
Locality kernels for sequential data and their applications to parse ranking.
Applied Intelligence, 31(1):81-88, 2009.
Regularized least-squares for parse ranking.
In A. Fazel Famili, Joost N. Kok, José Manuel Peña, Arno Siebes, and A. J. Feelders, editors, Proceedings of the Advances in Intelligent Data Analysis VI (IDA 2005), pages 464-474. Springer, 2005.
Information extraction from biomedical text: The biotext project.
In Margit Langemets and Priit Penjam, editors, Proceedings of the Second Baltic Conference on Human Language Technologies (HLT 2005), pages 131-136. Institute of Cybernetics, Tallinn University of Technology , 2005.
To obtain the source code of the algorithms used for conducting experiments in these papers please send me an email. Some of the publications can be downloaded from TUCS or Machine Learning Group repository.
Links
- CNCR I am involved in GENEUSS (GEne NEtworks Underlying Synaptic Signaling) project. The general aim of the this project is to apply machine learning techniques to construct a dynamic model for the presynaptic gene network underlying short-term plasticity in neuronal synapses. We develop a novel Bayesian perturbation approach, which combines a computational representation of the system with controlled genetic perturbations in a Bayesian framework.
- Machine Learning Group
- Student Projects
- Turku BioNLP Group
Software
Probabilistic Preference Learner/Ranker - ProbRank
-
The algorithm can learn a ranking function based on pairwise comparison data, that is, data about the ranking function values is provided in terms of pairwise comparisons at the given locations. This is accomplished in two ways: a) Approximating the marginal likelihood using expectation propagation and carrying out maximum likelihood procedure on the hyper-parameters. In this case the square exponential covariance function is used.
b) Considering ranking as a regression with Gaussian noise and Gaussian processes prior, given the score differences.
You can download Matlab implementation of the probabilistic preference learning models described in "Kernel principal component ranking: Robust ranking on noisy data". Please cite the paper in case you are using this code. Written by Botond Cseke.