Research projects
Complex Systems Computation Group - CoSCo
EU Network of Excellence in Pattern Analysis, Statistical Modelling and Computational Learning (PASCAL)
Period: 12/2003-2/2008
Researchers: Petri Myllymäki [Site Manager], Wray Buntine, Patrik BJ Floréen, Aapo Hyvärinen, Matti Kääriäinen, Jyrki Kivinen, Tei Laine, Taneli Mielikäinen, Vladimir Poroshin, Jorma Rissanen, Juho Rousu, Esko Ukkonen, Huizhen Yu, Janne Kataja, Rahul Katragadda, Jukka Kohonen, Jussi T Lindgren, Teemu T Roos, Tomi Silander, Abhishek Tripathi, Kimmo Valtonen, Hannes Wettig, Juan Carlos Borrás García, Petri Kontkanen, Tommi Mononen, Petteri Nurmi, Esa Pitkänen, Jukka Suomela, Ville Tuulos
Funding: EU
Pascal is a Network of Excellence on Pattern Analysis, Statistical Modeling and Computational Learning, funded by the EU and comprising 57 European research institutions. The Department of Computer Science at the University of Helsinki is one of the thirteen core sites of the network, and a representative from the University of Helsinki has a seat in the steering committee of the network. The basic idea of the network is to gather the foremost experts on statistical modelling and machine learning in Europe .
In 2007, the University of Helsinki node continued its active participation in the network in, among other things, the thematic programmes, Challenges, and the SIG on Information-Theoretic Modeling. Petri Myllymäki continued as a member of the steering committee of Pascal. Additional information: http://www.pascal-network.org/
SensorPlanet
Period: 4/2007-6/2007
Funding: Nokia
Researchers: Petri Myllymäki, Ville Tuulos, Tomi Silander, Jukka Perkiö
The SensorPlanet initiative, innovated and set up by the Nokia Research Center, aims at building an open global mobile device centric research platform for Wireless Sensor Network (WSN) research (akin to somewhat analogous platform for backbone network services called PlanetLabs). The distributed platform will provide the necessary infrastructure for world's top research labs to perform innovative research on wireless sensor networks, where the mobile devices can be seen both as gateways to the mesh sensor networks and also as sensor nodes themselves. This open innovation initiative will allow Nokia to collaborate with the best teams in the field around the world, and direct the academic Wireless Sensor Network research globally towards a mobile device centric innovation. More information about the SensorPlanet initiative can be found at http://www.sensorplanet.org/ .
In addition to the generic SensorPlanet open initiative, Nokia ran a Tekes funded research project, which supported local SensorPlanet-related research work in Finland , and the SensorPlanet project at University of Helsinki was part of Nokia's Tekes project work via subcontracting to the Cosco group.
Cognitive-Level Annotation using Latent Statistical Structure (CLASS)
Period: 01/2006-12/2008
Researchers: Petri Myllymäki, Ville Tuulos, Antti Tuominen, Tomi Silander, Mika Urtela
Funding: EU
Class will develop a basic cognitive ability for use in intelligent content analysis: the automatic discovery of content categories and attributes from unstructured content streams. The demonstrators will focus on object recognition and scene analysis in images and video with accompanying text streams. Autonomous learning will make recognition more adaptive and allow more general classes and much larger and more varied data sets to be handled.
Technically, the work will combine latent structure models and semi-supervised learning methods from machine learning with advanced visual descriptors from computer vision and state-of-the-art text analysis techniques. Three levels of abstraction will be studied: new individuals (specific people, objects, scenes, actions); new object classes and attributes; and hierarchical categories and relations between entities.
Class is an interdisciplinary project, combining six leading European research teams in visual recognition, text understanding & summarization, and machine learning.
CLASS home page: http://class.inrialpes.fr/
Cognitively Inspired Visual Interfaces for Representing Multidimensional Information (CIVI)
Period: 01/2005-12/2008
Researchers: Petri Myllymäki, Jussi Lahtinen, Petri Kontkanen
Funding: Academy of Finland
The CIVI project studies how to visualize the multidimensional information that is available to everyone through e.g. different search engines. On the one hand, the question is studied as a mathematical dimension reduction problem, on the other, as a challenge in perceptual psychology. This inter-disciplinary research is carried out in a two-university consortium, comprising the Cosco group at the University of Helsinki , led by Professor Petri Myllymäki, and Docent Ilpo Kojo's research group at the CKIR unit at Helsinki School of Economics.
Search-Ina-Box (SIB)
Period: 03/2003-06/2007
Researchers: Petri Myllymäki, Wray Buntine, Jussi Lahtinen, Jaakko Löfström, Jukka Perkiö, Vladimir Poroshin, Antti Tuominen, Ville Tuulos, Kimmo Valtonen
Funding: Tekes, Patentti- ja rekisterihallitus, Nokia, Wisane, M-Brain
The SIB project developed next-generation methods for semantic information retrieval and personification based on automatic analysis of large-scale information sources. These methods were integrated to a set of prototypes that were tested in different pilot environments, of which we can mention the following examples: The Aino search engine for the Finnish Internet (aino.hiit.fi), a semantic patent search engine (patent.hiit.fi), and a new search engine for the English Wikipedia. These ambitious pilots worked as a backbone for the research and inspired and guided the basic research done in the project. The most important pilots are public, and are based on open source software.
Since information retrieval will be the fundamental basic service required in future information networks, the potential applications of the SIB technology are numerous. The methods developed in the SIB project form a basic infrastructure for future web-based information-management systems, both for the closed intranets of companies and for open systems providing Internet information (such as Internet search engines).
There were three partners in the research consortium: the Department of Computer Science at the University of Helsinki /the Helsinki Institute for Information Technology HIIT (Professor Petri Myllymäki, the coordinator), the Department of Computer Sciences at the University of Tampere (Professor Kari-Jouko Räihä), and the Department of Health Policy and Management at the University of Kuopio (Professor Olli-Pekka Ryynänen). Additional information: http://cosco.hiit.fi/search/
MDL-Based Methods for Image Denoising (KUKOT)
Period: 1/2006-7/2008
Researchers: Petri Myllymäki, Jorma Rissanen, Teemu Roos, Hannes Wettig, Petri Kontkanen, Tomi SIlander, Tommi Mononen
Funding: Tekes
We can consider digital bit streams processed in the ICT sector as consisting of two overlapping parts, where one part is useful information and the other is useless noise. There is noise in all digital media; it is generated by the faults in original information sources (such as poor image resolution) and errors in signal transmission (such as disruptions in wireless communications or faults in hard drives). Noise can be filtered if the features of the source are known (in some degree at least), but it is very difficult to build general methods for denoising since they have to be able to construct adaptive models of random noise sources. The main problem with such adaptive modelling is the regularization of models; too complex (over-adaptive) models will interpret noise as part of the information and thus be rendered useless.
MDL (Minimum Description Length) is an information-theoretical framework developed by the father of arithmetic encoding, Jorma Rissanen. It provides an elegant solution for this problem. Unfortunately, the methods based on the MDL theory are often very challenging computationally. In 2007, the project team focused on studying how to implement MDL in a manner that is feasible for practical applications, and managed to develop computationally efficient methods suitable for tree-structured discrete Bayesian networks. In addition, the team developed two new variants of the NML criterion: sequential NML and factorized NML. The analysis of these new methods is still in progress.
The research consortium consists of two sub-groups: the Complex Systems Computation group at the Department of Computer Science at the University of Helsinki (Prof. Petri Myllymäki, the coordinator) and the Laboratory of Computational Technology at Helsinki University of Technology (Dr. Jukka Heikkonen). Additional information: http://ww.mdl-research.org
Probabilistic Methods for Microarray Data (PMMA)
Period: 1/2004-4/2008
Researchers: Petri Myllymäki, Jorma Rissanen, Teemu Roos, Hannes Wettig, Jussi Lahtinen, Tomi Silander, Petri Kontkanen
Funding: Tekes
The main objective of the research is to develop advanced methods for microarray data analysis. In particular the project focuses on the following research issues: denoising of microarray images, gene clustering and classification, and estimation of the reliability of the results.
The research consortium consists of three partners: The Laboratory of Computational Engineering (LCE) at Helsinki University of Technology (Academy research fellow Dr.tech Jukka Heikkonen, the coordinator), the Complex Systems Computation Research Group (CoSCo) at University of Helsinki (Prof. Petri Myllymäki) and the Institute of Biomedicine at University of Helsinki (Prof. Tomi Mäkelä).
In 2007 the project work continued in studying algorithms for finding the globally optimal Bayesian network, and the team showed that in contrast to what had been believed, the optimization problem is surprisingly sensitive to the value of the so called equivalent sample size parameter. The project also studied means for using histogram density functions for handling continuous variables