Suomeksi På svenska In English
University of Helsinki Department of Computer Science
 

Annual report 2006

Helsinki Institute for Information Technology HIIT/BRU

The Helsinki Institute for Information Technology (HIIT) is a joint research unit of the University of Helsinki and Helsinki University of Technology. The unit aims at significantly enhancing strategic research in the field of computer science, while improving the co-operation between the University of Helsinki and the Helsinki University of Technology. HIIT consists of two units: the Advanced Research Unit (headed by Martti Mäntylä), which cooperates with the industries and functions primarily at the University of Technology , and the Basic Research Unit (headed by Esko Ukkonen).

Some fifty researchers work in BRU. The unit is mostly located at the Department of Computer Science at the Kumpula campus. Part of it operates at Helsinki University of Technology in Otaniemi. The Basic Research Unit (BRU) carries out first-rate basic research in the field of computer science, in close cooperation with end-users in other disciplines and in the industries. Its main research areas are theory and applications in data analysis, adaptive computing and neuroinformatics.

Year 2006 saw continued use of the ContextPhone software, which has proved to be a remarkable innovation. The further development of this software has partially been transferred to a company established by the researchers.

In the field of data-analysis algorithms several studies were completed, in connection with e.g. segmentation of data sets, inner dimensions and order reconstruction. The results were applied in fields like genetics and palaeoecology. A collaboration was started with researchers at MIT (Professor Tommi Jaakkola).

The neuroinformatics group continued developing the LiNGAM (Linear non-gaussian acyclic model) method. This method finds causal relations between linear data variables. The method was enhanced for some very important practical cases: 1) some variables are not observed, or 2) the data is hidden. The group managed to discover new interesting connections between their own methods and those of Geoffrey Hinton's ( Toronto ) group.

Contact persons: Professor Esko Ukkonen, Academy Professor Heikki Mannila

Webpage: http://www.cs.helsinki.fi/hiit_bru/ , http://www.hiit.fi/

Projects

Networks and architecture for proactive systems - algorithmics
Software Platform and Component Environment for yoU
MobiLife - Mobile Life
Advanced data analysis in vision research
Statistical modelling of image and video data
Independent component analysis and its extensions
Computational methods for the analysis of genome structure and function in mammals
APRIL II
Context-awareness with the help of user location data mining
Searching for predisposition genes in case control data

 

Publications

Buntine, W. & Jakulin, A.: Discrete component analysis. SLSFS 2005: Subspace, latent structure and feature selection. - Berlin : Springer cop. 2006. p. 1-33.

FloreŽen, P. & Kukkonen, J. & Lagerspetz, E. & Nurmi, P. & Suomela, J.: BeTelGeuse : tool for context data gathering via BluetoothLinkit. Proceedings of the Second Workshop on Context Awareness for Proactive Systems. - Kassel : Kassel University Press 2006. p. 137-139.

Fortelius, M. & Gionis, A. & Jernvall, J. & Mannila, H.: Spectral ordering and biochronology of European fossil mammals. Paleobiology. Chicago , IL . 32 (2006): 2, s. 206-214.

Kaski, P. & Östergård, P.R.J. : Classification algorithms for codes and designs. Berlin : Springer, 2006, 412 p. Series: Algorithms and computation in mathematics.

Raento, M.: The data subject's right of access and to be informed in Finland : an experimental study. International journal of law and information technology. 14 (2006) : 3, p. 390-409.

Shimizu , S. & Hyvärinen, A. & Hoyer, P.O. & Kano, Yutaka: Finding a causal ordering via independent component analysis. Computational statistics & data analysis. - Amsterdam . 50 (2006), p.