Index theory on the Miščenko bundle (Q832820): Difference between revisions

This paper was very well written and can be extended as a textbook smoothly. Among many interesting results in this paper, the authors proved the connection of Baum-Connes index map and the Miščenko-Fomenko index map, and a generalized Atiyah's \(L^2\)-index theorem. \[ \begin{tikzcd} KK^G_*(C_0(\tilde{X}),\mathbb{C})\arrow[dr, "\mu_{\tilde{X}}"] \arrow[r, "J_{\tilde{X}}"] & KK_*(C(X),\mathbb{C})\arrow[d, "\eta_{\tilde{X}}"] \arrow[dr, "\operatorname{ind}"] \\ & KK_*(\mathbb{C},C^*_r(G))\arrow[r, "\phi_*"] & \mathbb{R} \end{tikzcd} \] With the help of these results, the authors gave a new proof of the result that ``the surjectivity of the Baum-Connes assembly map implies the Kadison-Kaplansky idempotent conjecture in the torsion-free case''. Without doubts, all results in this paper have their own independent interests. In the introduction part, the authors provided a list of parts which they need to replace in traditional proof in order to give a ``self-contained and topological'' proof of the implication. In the Section 1, the authors set up all necessary notations and defined the Baum-Connes assembly map, the Miščenko-Fomenko index map and the dual Green-Julg map which linked the previous two index maps. Then the authors mentioned two main theorems of the paper: \((A)\) the dual Green-Julg map is an isomorphism and \((B)\) the Miščenko-Fomenko index, as a generalization of \(L^2\) index of Atiyah. In the second section, the authors prove Theorem A by providing explicit inverses of Green-Julg's map and the constructions of equivalence of Miščenko line bundle. So the Baum-Connes index map is linked to the Miščenko-Fomenko index map. In the Section 3, an explicit Chern character was constructed and an isomorphism theorem was proved for \(II_1\) factors. This section was natural, but technical, and it was introduced here to avoid the usage of the geometric \(K\)-homology, as promised in the introduction part. In the final section, Theorem B was proved. As we can see that the authors constructed and proved theorems in details. Some parts were very technical. It is better to read the final section, the proof of Theorem 4.1 first, especially the following computation: \begin{align*} \operatorname{ind}_A(x)&=\operatorname{ind}_B(x)=\lambda_*([p_B]\hat{\otimes}_{C(X,B)}\tau_B(x))\\ &=\lambda_*([i(q)]\hat{\otimes}_{C(X,B)}\tau_B(X))=\phi(q)\cdot\operatorname{ind}(x). \end{align*} Then all possible doubts may appear and then it will be more motivated to understand previous sections and lemmas.