Discovery group: Data Mining for Pattern and Link Discovery
The Discovery group develops novel methods and tools for data mining and computational creativity. Our focus is on algorithmic methods for discovering links and patterns in data, and recently also on their use in creative systems. Application areas range from link discovery in bioinformatics to computational generation of poetry.
Data mining and graph mining. Methodologically, we focus on analysis and exploration methods for weighted graphs. We identify relevant computational problems, develop new concepts and algorithms, and apply them. We have recently developed a range of novel methods to simplify large networks into simpler ones or for extracting relevant information from them. These methods allow more efficient and user-friendly analysis of social networks, biological networks, etc. See Publications page for results in the area.
Biological networks. The more applied line of this research has produced Biomine, a search engine prototype that integrates and indexes data from several publicly available biological databases. Biomine presents the data as a weighted graph, and its query tools aid explorative discovery of non-trivial connections between biological entities, such as genes and phenotypes. See http://biomine.cs.helsinki.fi/.
Computational creativity. A new focus area is computational creativity, interesting on its own right but also as an application area for data mining methods. We work on verbal creativity, i.e., computational poetry and humour, and also on automatic composition of music. We are developing novel methods that minimize the need for manually coded or language-specific knowledge.
Artistic results. We have built a revolutionary data musicalization service that composes a novel piece of music from the users' sleep measurements. See sleepmusicalization.net for an implementation, example songs, and more information. (This work won the IDA 2012 Frontier Prize for 'the most novel and visionary contribution' of the conference!) Poems by P. O. Eticus-Apparatus, a program from the Discovery group, have been used as the text of the art book Tee se kotona. Kiitos. by artist Heikki Paakkanen.
The group works jointly under the Department of Computer Science at University of Helsinki, and The Helsinki Institute for Information Technology (HIIT). We constitute a part of Algodan (Algorithmic Data Analysis), a national Centre of Excellence in research.
A sample of media coverage of Discovery group since 2011:
(Finnish TV news) (Finnish) (English) (English) (English) (German) (German) (French) 雷锋网官方博客 (Chinese) (Greek) (Finnish radio) (English) (English)
See sleepmusicalization.net for many more countries and languages