Investigation of dependencies between multiple data sources allows the discovery of regularities and interactions that are not seen in individual data sets. The increasing availability of co-occurring measurement data in computational biology, social sciences, and in other domains emphasizes the need for practical implementations of general-purpose dependency modeling algorithms.

The project collects various dependency modeling approaches into a unified toolbox. The techniques for the discovery and analysis of statistical dependencies are based on well-established models such as probabilistic canonical correlation analysis and multi-task learning whose applicability has been demonstrated in previous case studies.