Some News

Presentation at the Seminars of the Intelligent Decision Support Systems Lab:
With Kalina Jasińska and Marek Wydmuch we gave a talk for the IDSS lab about probabilistic label trees: [pdf].

A lecture given during the Predoc Summer School on Learning Systems on Monday, July 3, 2017, at ETH Zürich.
Slides from the lecture: [pdf].

A lecture at Adam Mickiewicz University in Poznań
Slides from my talk on extreme multilabel classification given in a series of open lectures on "Mutlivariate statistical methods for engineering" [pdf].

The 3year project "Consistent and scalable learning algorithms for structured output prediction" financed by the National Science Centre (NCN) has been finished.
The main achievements of the project are:
 A new extreme classficiation algorithm for multilabel problems, referred to as Probabilistic Label Trees [icml paper] [code].
 A theoretical analysis of complex performance measures [MLJ paper]
 An analysis of Fagin' threshold algorithm in a wide spectrum of machine learning problems [DAMI paper]
 A theoretical analysis of probabilistic classifier trees [ECML PKDD paper]
 An online algorithm for Fmeasure maximization [NIPS paper]

An invited talk at the TFML 2017 conference
Slides from my talk on Extreme ZeroShot Learning [pdf].

The Best Paper Award for a paper with Wojciech Kotłowski at the Asian Conference on Machine Learning 2015:
The paper, entitled Surrogate regret bounds for generalized classification
performance metrics, can be found here

Lectures on decisiontheoretic machine learning at Poznań University of Technology, Summer 2015:
Lecture is given with Wojtek Kotłowski. The slides from the lectures are available here:

An Open Position for a Phd Student in Machine Learning at Poznań University of Technology:
There is an open position for a PhD Student in the field of Machine Learning in the Laboratory of Intelligent Decision Support Systems at Poznań University of Technology.
The position is funded by the National Science Centre (NCN). The project name is "Consistent and scalable learning algorithms for structured output prediction."
The deadline for submitting applications is September 30, 2014.
For more information see here (in Polish).

A talk on label tree structures for efficient classification, 14 July 2014:
I gave a talk on label tree structures for efficient multiclass and multilabel classification at IFORS 2014 in the stream on Preference Learning organized by Roman Słowiński, Salvatore Greco, Willem Waegeman and me. The slides from the talk are available here.

Series of lectures on decisiontheoretic machine learning, from 30 June to 4 July 2014:
With Wojtek Kotłowski we gave a series of lectures on decisiontheoretic machine learning at the summer school organized by the Institute of Computer Science of the Polish Academy of Sciences. The slides from the lectures are available here:

Presentation at the Seminars of the Institute of Computing Science, Poznań University of Technology, 27 May 2014:
I gave a presentation on multilabelclassification at the Seminars of the Institute of Computing Science. The slides from the talk are available here.

Three talks on machine learning at Poznań University of Technology, 25 March 2014:
Eyke Hüllermeier will give a talk entitled "Preference Learning: Methods and Algorithms for Ranking Problems" at the Seminar of the Institute of Computing Science on Tuesday, 25 March 2014, 11:1512:15. Two others talks of Weiwei Cheng and Robert BusaFekete are scheduled for the Seminar of the IDSS Labs which starts at 13:30. The official announcement can be found here (in Polish)

Meeting of the Polish Special Interest Group in Machine Learning, 1213 November 2013:
Meeting of the Polish Special Interest Group in Machine Learning took place in Warsaw on 1213 November 2013. My talk given there was a shorter version of my DS/ALT and ICML tutorials. The slides from the talk are available here.

Tutorial on MultiTarget Prediction at ALT/DS 2013:
I gave a tutorial on MultiTarget Prediction at the Discovery Science 2013 conference (colocated with Algorithmic Learning Theory 2013).The slides from the talk are available here.

An Open Position for a Master Student in Machine Learning at Poznań University of Technology:
There is an open position for a Master Student in the field of Machine Learning in the Laboratory of Intelligent Decision Support Systems at Poznań University of Technology.
The position is funded by the Foundation for Polish Science (FNP) under the Homing Plus program. The project name is "Collective Learning and Inference in MultiTarget Prediction Problems."
The position is scheduled for 6 months and starts in November, 2013. The deadline for submitting applications is October 1, 2013.
For more information see here (in Polish).

Tutorial on MultiTarget Prediction at ICML 2013:
We gave a tutorial on MultiTarget Prediction at ICML 2013. For more details check here.
The videos from the tutorial are available on techtalks.tv.
A similar talk will also be presented at the Discovery Science 2013 conference (colocated with Algorithmic Learning Theory 2013).
Selected publications

Extreme FMeasure Maximization using Sparse Probability Estimates
Kalina Jasinska, Krzysztof Dembczyński, Robert BusaFekete, Karlson Pfannschmidt, Timo Klerx, Eyke Hüllermeier
Proceedings of the 33rd International Conference on Machine
Learning, New York, NY, USA, 2016. JMLR: W&CP, 48 14351444, 2016

On the Bayesoptimality of Fmeasure maximizers
Krzysztof Dembczyński, Willem Waegeman, Arkadiusz Jachnik, Weiwei Cheng, Eyke Hüllermeier
Journal of Machine Learning Research, 15 33333388, 2014

On loss minimization and label dependence in multilabel classification
Willem Waegeman, Krzysztof Dembczyński, Weiwei Cheng, Eyke Hüllermeier
Machine Learning, 88 545, 2013
Listed in Notable Computing Books and Articles 2012 by ACM Computing Reviews

Learning monotone nonlinear models
using the Choquet integral
Ali Fallah Tehrani, Weiwei Cheng, Krzysztof Dembczynski, Eyke Hüllermeier
Machine Learning, 89 183211, 2013

ENDER: a statistical framework for boosting decision rules
Krzysztof Dembczyński, Wojciech Kotłowski, Roman Słowiński
Data Mining and Knowledge Discovery, 21 5290, 2010

Optimizing the Fmeasure in multilabel classification: Plugin rule approach versus structured loss minimization
Krzysztof Dembczyński, Arkadiusz Jachnik, Wojcieh Kotlowski, Willem Waegeman, Eyke Hüllermeier
International Conference on Machine Learning (ICML 2013) 2013

An analysis of chaining in multilabel classification
Krzysztof Dembczyński, Willem Waegeman, Eyke Hüllermeier
European Conference on Artificial Intelligence (ECAI 2012) 294299, 2012
Best Paper Award

Consistent multilabel ranking through univariate losses
Krzysztof Dembczyński, Wojciech Kotłowski, Eyke Hüllermeier
International Conference on Machine Learning (ICML 2012) 2012

An exact algorithm for Fmeasure maximization
Krzysztof Dembczyński, Willem Waegeman, Weiwei Cheng, Eyke Hüllermeier
Advances in Neural Information Processing Systems 25 (NIPS 2011) 2011

Bipartite ranking through minimization of univariate loss
Krzysztof Dembczyński, Wojciech Kotłowski, Eyke Hüllermeier
International Conference on Machine Learning (ICML 2011) 2011

Bayes optimal multilabel classification via probabilistic classifier chains
Krzysztof Dembczyński, Weiwei Cheng, Eyke Hüllermeier
International Conference on Machine Learning (ICML 2010) 2010
You can find more my publications on Google Scholar.
Lecturing
Software

Extreme MultiLabel Classification (XMLC)
Extreme MultiLabel Classification (XMLC) deals with multilabel problems with hundreds of thousands of labels. Implementation of Probabilistic Label Trees (PLT) suited for this kind of problems can be found here: https://github.com/busarobi/XMLC.

Probabilistic Classifier Chains:
Probabistitic Classifier Chains (PCC) are a learning method for multilabel classification problems. You can find the code in the GitHub repository: https://github.com/multilabelclassification/PCC.

Ensembles of Decision Rules:
There are two fast and accurate boosting algorithms for learning decision rule models. For multiclass problems use MLRules (Maximum Likelihood Rule Ensembles), while for regression problems use RegEnder (regression ensemble of decision rule).
