We describe a statistical relational learning framework called Gaussian Logic capable to work efficiently with combinations of relational and numerical data. The framework assumes that, for a fixed re
