This is a sub-entry for gerbe.
For related entries see
Gerbes give a nice way to group together bundle data on a smooth manifold, but gerbes also naturally define degree two cohomology. Thus the idea of using gerbes in differential geometry is to have a nice language that relates geometric concepts such as connections and curvature to cohomological classifications.
In addition, one can use the analogies above that are made precise with gerbes to define other new concepts such as 3-curvature and “local” fiberwise connections.
Let be a smooth manifold. A Dixmier-Douady sheaf of groupoids over is a - gerbe on where is the sheaf of smooth -valued functions (not to be confused with the constant sheaf ).
We define to be the term in the exponential sequence on : .
Taking the associated sequence in cohomology to the exponential sequence gives us an isomorphism .
We have a canonical isomorphism between the group of equivalence classes of Dixmier-Douady sheaves of groupoids over (basically by the definition of -gerbe) and .
Matt: Actually, a slight issue has arisen. Most of the things I thought would go here actually already appear in other places even though they aren’t grouped as coming from the same idea.
For instance, bundle gerbe contains the geometric interpretation of . Also, 3-curvature and fiber-wise connections occur at connection on a bundle gerbe. Although I think there is still a lot to say, I’m not convinced that “gerbe (in differential geometry)” is necessary anymore
Jean-Luc Brylinski Loop Spaces, Characteristic Classes, and Geometric Quantization
I. Moerdijk, Introduction to the language of stacks and gerbes (arXiv)
Larry Breen, Notes on 1- and 2-gerbes (arXiv)
Further references are given in the other entries on gerbes.