Discovery group: Data Mining for Pattern and Link Discovery
We develop novel methods and tools for pattern and link discovery. Our focus is on structured and heterogeneous data, such as graphs, and also on sequences. The main applications are in bioinformatics, genetics, and in ubiquitous computing, in tight collaboration with applied scientists and companies. The group is part of the Algodan Centre of Excellence.
Our current emphasis is on analysis and link discovery in weighted (biological) graphs. Our new methods are used to implement the Biomine search engine for biological data, as well as for visualisation of the search results. In the past year, a focal theme was extraction of a relevant subgraph from a large graph.
The group introduced, together with universities of Freiburg and Leuven , a new probabilistic Prolog, ProbLog, and an efficient implementation for it. ProbLog combines logical and probabilistic reasoning in a simple framework. It can be applied on weighted graphs, among others.
The ContextPhone platform that was developed in the group for context-aware mobile applications, reached a new level of impact when Google bought Jaiku in October last year. ContextPhone was the technological basis for Jaiku, a popular activity stream and presence sharing service since 2006.
Contact person: Professor Hannu Toivonen
Homepage: http://www.cs.helsinki.fi/research/discovery
Projects
Biomine
ProbLog
Context
Selected publications
L. De Raedt, A. Kimmig, H. Toivonen: ProbLog: A Probabilistic Prolog and its Application in Link Discovery. Twentieth International Joint Conference on Artificial Intelligence (IJCAI-07), 2468-2473, Hyderabad , India , January 2007.
A. Oulasvirta, R. Petit, M. Raento, S. Tiitta: Interpreting and Acting on Mobile Awareness Cues. Human-Computer Interaction 22:97-135. 2007.
P. Hintsanen: The Most Reliable Subgraph Problem. 11th European Conference on Principles and Practice of Knowledge Discovery in Databases (PKDD), 471-478, Warsaw , Poland , September 2007.
L. De Raedt, A. Kimmig, H. Toivonen: Probabilistic Explanation Based Learning. 18th European Conference on Machine Learning (ECML), 176-187, Warsaw , Poland , September 2007. Winner of the ECML-07 Best Paper Award.
N. Haiminen, A. Gionis, K. Laasonen: Algorithms for unimodal segmentation with applications to unimodality detection. Knowledge and Information Systems 14:1, 39-57. 2007.