Social Networks and Information Flow, LastFM

    The goal of this project is to understand information flow in virtual social networks. We try to track and model the dissemination of new songs (new information) within LastFM user network.

    Gaussian Mixture Modeling (GMM) and Fisher vector Toolkit

    Gaussian Mixture Modeling (GMM) and Fisher vector[1] C/C++ source code for image classification on GPUs and x86 CPUs

    Rev. 0.94 (see README in the compressed file) Download

    CLIFF alpha Rev. 0.11 (see README in the compressed file) Download

    Text Mining for Criminal Investigation Offices

    2006-2008. Government funded project to develop text and video mining tools for crime investigation. Collaboration with Tamás Szirányi's team for video event detection, modeling, classification and annotation.

    VAST 2011

    Our team has won "Outstanding Integration of Computational and Visual
    Methods" award on IEEE Visual Analytics Science and Technology Challenge 2011.

    We have developed a user-friendly tool, City Sentinel, that supports analysis of geospatial microblog data (short text messages with timestamp and geographical coordinates).
    With our tool, we could easily answer the questions about the spread of a fictional disease within a city. The solution integrated data mining and text mining algorithms with a usable graphical interface.

    Reticular Alignment: algorithm for multiple sequence alignment

    Reticular Alignment is our new method for for multiple sequence alignment. Unlike previous corner-cutting methods, our approach does not define a compact part of the dynamic programming table. Instead, it defines a set of optimal and suboptimal alignments at each step during the progressive alignment. The set of alignments are represented with a network to store them and use them during the progressive alignment in an efficient way.

    MCMC Network: Graphical Interface for Bayesian Analysis of Metabolic Networks

    Metabolic network Metabolic networks are under continuous evolution. Most organism have a common set of reactions as a part of their metabolic networks that relate to essential processes but a large proportion of the reactions present in different organisms are specific to the needs of individual organisms or tissues. The regions of metabolic networks corresponding to these non-essential reactions are under continuous evolution.


    Strategic Crime & Immigration Information System

    The project aids combating organized crime with focus on people trafficking and smuggling. We developed information management techniques enabling information fusion and sharing within a secure information infrastructure, tools to assist decision making in order to predict and analyze people trafficking and smuggling sources, events and links to organized crime. We utilized and enhanced state of the art products with new innovative technologies to improve the ability of European agencies to fight organized crime.

    Small molecule similarity modeling, Richter

    Small molecule similarity modeling and clustering for Richter Richter Gedeon Ltd. 2004-2006.

    Molecules are easy to be represented by labeled graphs. Unfortunately, this representation is not very suitable for computing similarity between molecules.
    Therefore, fingerprint models are applied to solve this problem much more efficiently.


    EU funded project to develop data mining tools for in silico drug discovery. Results: small molecule fingerprint model.