Seminar: Constraint Solving Meets Machine Learning and Data Mining (Spring 2013)


Constraint programming, maching learning, and data mining are today well-established and thriving research fields within computer science. Each of the fields have contributed fundamental techniques and algorithmic solutions that are today routinely applied for addressing hard computational problems in various real-world contexts, including both industrial and scientific applications.

However, these fields have developed somewhat independently of each other. Especially, the possibilities of exploiting the highly optimized constraint solving technology available today in providing generic and efficient solutions to various machine learning and data mining problems, such as different kinds of classification, structure-learning, probabilistic reasoning, and pattern mining tasks, have only recently been realized. In addition, machine learning techniques are starting to be employed in further improving constraint solvers e.g. by building intelligent constraint solver portfolios based on model-based selection.

The aim of this seminar is to provide a snapshot into the current state of forefront research in the intersection of constraint solving, machine learning and data mining, by studying recently published research articles. Each student will select one or more articles based on which (s)he will write a seminar report and give a seminar presentation during two workshop days to be held during May 6-14.

This is a very topical area of research, and possibilities of extending the seminar work into an MSc thesis may be available (consult the teacher) within the COIN Centre of Excellence in Computational Inference Research. For more background on the seminar topic, see e.g. Dagstuhl seminar on Constraint Programming meets Machine Learning and Data Mining, and CoCoMile 2012: First workshop on COmbining COnstraint solving with MIning and LEarning.

Course Information

Course Requirements

See also the introductory slides from the first meeting on March 13.


Keep in mind that the written report and the oral presentation have partially different goals.

Your report should look like a typical scientific article. However, it will not contain any new scientific results, just a survey of previously published work.

Everything that you write in your seminar report must be written in your own words. Use appropriate citations to make it clear which result comes from which publication.

All illustrations and examples should be your own original work. There is usually no need to use the same example as what was used in the original publications; by constructing your own examples you are also forcing yourself to actually understand the details of the result.

(If it is absolutely necessary to reuse illustrations or examples from other sources, make it very explicit in your report, with appropriate citations, and make sure you are not violating anyone else's copyrights. Remember that simply re-drawing a figure does not make it your own work.)

Important: Make sure that your seminar report (and presentation) shows that you have actually understood everything, it is completely clear to you, and now you are doing your best to explain the key ideas to others. Do not follow the structure of the original articles; come up with your own presentation that is much better than the original version! Ideally, pick ideas from several papers and synthesise.


The seminar meetings start on the hour, not 15 minutes past. You must arrive at least 15 minutes before, in order to setup your presentation!

Date Presenter Topic  Opponent
Tuesday May 7 14-18 in C220 Jeremias BergG1Timo Koivisto
Mikko SysikaskiG2Jeremias Berg
Timo KoivistoG3Mikko Sysikaski
Quan NguyenD1Aleksi Hartikainen
Aleksi HartikainenD2Jussi Kokkala
Jussi KokkalaD4Quan Nguyen
Tuesday May 14 12-16 in B119   Alejandro Sanchez Guinea  A1
Lari RaskuA3Simo Linkola
Simo LinkolaA2Lari Rasku
Liye HeD3Matthew Pierce
Matthew PierceB1Kustaa Kangas
Kustaa KangasB2Liye He


Below are listed possible topics for giving a presentation and writing a report. Each student chooses one topic, and one topic can be chosen only by one student. The provided articles for each topic serve as the main references, but it is a plus if you independently consult additional suitable references. The topics will be assigned during the first seminar meeting (introduction to the topic presented by the teacher) on a first-come-first-served basis. You can also reserve a topic before the first meeting by contacting the teacher.

It is also possible suggest a topic outside this list; if you wish to suggest one, please discuss details with the teacher.

A. Data Mining using constraint solvers / knowledge compilation B. Using machine learning to configure / speed-up constraint solving C. Clustering with constraints D. Learning Bayesian networks using constraint solvers / heuristic search E. Learning causal models using constraint solving F. Model counting and probabilistic inference G. Further topics