The paper tackles a classical problem in algebraic geometry: understanding the relationship between two fundamental ways of measuring geometric complexity on a space. One approach uses algebraic K-theory, which tracks vector bundles (roughly, smoothly varying families of vector spaces) over a space. The other uses Chow groups, which track algebraic cycles, meaning formal combinations of subvarieties. The famous Riemann-Roch theorem, in its various incarnations, provides a precise dictionary between these two perspectives. The challenge here is extending this dictionary to singular spaces, where the geometry is not smooth and standard tools break down.
The author works with a modified version of the Chow group of zero-dimensional cycles, originally introduced in a 1985 collaboration with Weibel, that is designed to handle singularities carefully. The key idea is to separate the truly singular part of the space from the rest, treating them differently. Using this modified group, the paper constructs two maps: one going from K-theory to the modified Chow group, and another going in the reverse direction. These maps are analogous to familiar constructions in the smooth setting but require substantial new ideas to define correctly in the presence of singularities.
The main result is a Riemann-Roch theorem that describes precisely what happens when you compose these two maps in either order. In the smooth setting, such compositions are controlled by simple numerical factors coming from characteristic classes, and the paper shows an analogous story holds here despite the singular geometry. This extends and unifies earlier results, providing a coherent framework for doing intersection theory and computing invariants of singular algebraic varieties, which appear naturally throughout geometry and number theory.