http://lanl.arxiv.org/abs/hep-th/0411073 Robbert Dijkgraaf, Sergei Gukov, Andrew Neitzke, Cumrun Vafa See page 11 section 3.3 "4D 2-Form Gravity" The authors are well-known. Dijkgraaf is in Amsterdam and the others are at Harvard. "4D 2-form gravity" serves as a general pointer in the direction of LQG and its relatives. The connection variable on which LQG is based (the configuration set consists of connections embodying possible geometries) is technically a 2-form. LQG is a 4D connection-based model of gravity so it can be termed a "4D 2-form gravity" there are various other N-form-type gravity models with other dimensionality, like 3D and 7D etc. They dont all use 2-forms, so let's say N-form to keep it general. what Dijkgraaf et al try to do is to include all these various-dimensioned N-form-based models-----at least in a loose general way----within a theoretical framework they call "Topological M-Theory". As they indicate the connections are tentative----a bit fuzzy---and leave a lot of details to work out. The contact with LQG is not with something you can get your hands on and calculate with, but instead with what they call "topological LQG"-----something with the right general look AFAICS but not copying at the level of specifics.