I know things are already too complicated, but perhaps in a way they are too simple. Edges in hypergraphs are directed, which means they are 2-valued. Why not allow more general edges, like n-valued, continuous, vector/tensor valued, etc? I doubt particles are possible without topological singularities, and so far (though I haven't investigated this much) it doesn't seem to be possible--I am not impressed by "forbidden minors".