Computer Science

Algorithmics and Computational Intelligence Group (ACI)

The research of the laboratory is centered around techniques and methods for algorithm design and computational intelligence, with the emphasis on both theory and applications. The foundations of the research are machine learning, probabilistic inference, discrete mathematics and theoretical computer science. In particular, the research of kernel methods, Bayesian analysis, probabilistic and information-theoretical modeling, combinatorial algorithms and intelligent systems has been pursued. The research of the laboratory is based on the long tradition of combining basic research on algorithm development and analysis with active cooperation with companies and academic partners on solving real-life problems by the use of combinatorial optimization and latest techniques on computational intelligence methods.


Selected research topics on algorithmics and computational intelligence:

  • Classification and regression methods
  • Clustering methods
  • Combinatorial algorithms and applications
  • Cross-validation methods
  • Data compression
  • Feature selection methods for high-dimensional data
  • Industrial algorithms
  • Information retrieval
  • Information theoretic methods
  • Multi-task and transfer learning
  • Preference learning and ranking
  • Probabilistic Bayesian methods
  • String algorithms
  • Tensor product kernels for pairwise learning


Selected application projects:

  • Algorithmic methods in circuit board assembly manufacturing
  • Algorithms for embedded systems
  • Arctic soil analysis
  • Automated plant identification
  • Automated summarization of clinical narratives
  • Drug-target interaction prediction
  • Functionally related enzyme identification
  • Genomic data analysis
  • Magnetic resonance image analysis
  • Machine learning software development
  • Medical data analysis
  • Video compression


Associated researchers: Jukka Heikkonen, Tapio Pahikkala, Antti Airola, Lasse Bergroth, Juho Heimonen, Harri Merisaari, Ileana Montoya Perez, Parisa Movahedi, Juha Mäkilä, Pekka Naula, Olli Nevalainen, Paavo Nevalainen, Sebastian Okser, Sami Pietilä, Jonne Pohjankukka, Csaba Ráduly-Baka, Jukka Teuhola, Jussi Toivonen


Selected collaborators: Auria BioBank, FIMM Institute for Molecular Medicine , Geological Survey of Finland , Ghent University , Mentor Graphics Finland, Natural Resources Institute Finland, Oldenburg University, Szeged University, Turku PET Centre


News 2015:


IEEE Computational Intelligence Society has awarded a joint research article by reasearchers at the University of Turku, Department of Future Technologies. The article is a result of a long standing collaboration with researchers from the KERMIT research unit at the University of Ghent.


The IEEE CIS Outstanding TFS Paper award is worth one thousand dollars. The award was given for the article A kernel-based framework for learning graded relations from data. The research was done jointly by M. Waegeman, T. Pahikkala, A. Airola, T. Salakoski, M. Stock and B. De Baets. The article appeared in July 2012 in the IEEE Transaction on Fuzzy Systems, which is a top-tier journal in the area of computational intelligence. The award will be given in spring 2015 in the IEEE Congress on Evolutionary Computation in Japan.