Institute of Computing Science, Poznań University of Technology


Collective Learning and Inference in Multi-Target Prediction Problems

The project concerns problems in which output domains have complex structure, not limited to a single output variable. An example of such a problem is multi-label classification. The project is funded by Foundation for Polish Science (FNP) under the Homing Plus program, and the project leader is Krzysztof Dembczyński. The project consists of two parts:

  • The first part of the project concerns theoretical analysis of CLI methods with a particular emphasis on modeling the dependence between outputs/targets.
  • The second part is devoted to an application of CLI methods to a real problem taken from computational chemistry, namely to molecular docking.

Project Site >> (under construction…)
Resources >>