Welcome to MathOverflow

Question

4

2

In classical geometry the calculation of the Chern classes of a vector bundle using a connection is independent of the choice of connection. Does any such result hold for projective modules in non-commutative geometry?

edited Nov 18 2009 at 0:31

Scott Morrison♦
4,460●4●26●44

asked Nov 17 2009 at 23:47

John McCarthy
686●3●10

Mariano Suárez-Alvarez · Accepted Answer · 2009-11-18 00:13:16Z UTC

7

Yes.

You can see the construction in detail, for example, in Max Karoubi's ‘Homologie cyclique et $K$-theorie’ (Asterisque 149, SMF; you can get this from his web page), where he constructs the Chern classes $K_0(A)\to H(A)$ using connections much à Chern-Weyl (Here $H(A)$ is the non-commutative de Rham theory, or one of the various cyclic homologies of $A$) He also constructs higher Chern classes on the higher algebraic $K$-theory by a similar procedure. The book by Loday on cyclic homology also covers this.

edited Nov 18 2009 at 0:13

answered Nov 17 2009 at 23:58

Mariano Suárez-Alvarez
30.6k●7●68●153

	This leaves me a little confused. I am familiar with noncommutative vector bundles (ie projective modules) endowed with differential calculi and non-commutative connections, and then calculating the Chern Classes from this information. On the other hand, I know what cyclic cohomology is and why it can be viewed as a non-commutative de Rham cohomology. What I don't really understand is how the two link up: the former is more general than the latter? – John McCarthy Nov 18 2009 at 1:01
	... and if so, then Karoubi/Loday only prove the result in a special case? – John McCarthy Nov 18 2009 at 1:04
	There are maps connecting non commutative de Rham cohomology and cyclic homology (in fact, there is an isomorphism between $H_{\mathrm{dR}}(A)$ and the kernel of the map $B:HC_\bullet(A)\to HH_{\bullet+1}(A)$ appearing in the Connes long exact sequence---here $HH$ is Hochschild homology), and constructions of the Chern character from $K_0(A)$ to both $H_{\mathrm{dR}}(A)$ and $HC(A)$ which are compatible with those maps. A non trivial part of Max's Asterisque is spent in checking lots of such compatibilities. – Mariano Suárez-Alvarez Nov 18 2009 at 1:08
	Oh, when I say non commutative de Rham theory I mean something not really involving 'differential calculi' (or rather involving only the universal one): take $\Omega(A)$ to be the kernel of the multiplication map $A\otimes A\to A$, which is an $A$-bimodule, and let $\Omega^\bullet(A)$ be the tensor algebra over $A$ of $\Omega(A)$, which is a differential graded algebra. Now let $\Omega^\bullet(A)_{\mathrm{ab}}=\Omega^\bullet(A)/[\Omega^\bullet(A),\Omega^\bullet(A)]$ be the 'abelianization', which is a complex, whose cohomology is the non commutative de Rham cohomology I had in mind. – Mariano Suárez-Alvarez Nov 18 2009 at 1:15

Does the non-commutative Chern class depend on the choice of connection?

Remember to vote up questions/answers you find interesting or helpful (requires 15 reputation points)

1 Answer

You can accept an answer to one of your own questions by clicking the check mark next to it. This awards 15 reputation points to the person who answered and 2 reputation points to you.