Algebraic K-theory is a sophisticated tool that attaches sequences of algebraic invariants, called K-groups, to geometric spaces. One natural question is whether these invariants can detect how "nice" a space is geometrically. A space is called K-m-regular if a certain map between its K-groups behaves as well as it does for smooth spaces. Smooth spaces, meaning those without any corners, cusps, or other singular points, are always K-m-regular for every m. The question explored here is whether the reverse is true: if a space is K-m-regular, must it be smooth?
The paper focuses on a particular class of mildly singular spaces called finite quotient singularities, which arise by taking a smooth space and identifying points that are related by a finite group of symmetries. The main positive result is that for such spaces, being K-1-regular is enough to guarantee smoothness, provided the singular points are not spread out in a complicated way (for instance, if the space is affine or has only isolated singular points). This gives a clean and practical criterion for detecting smoothness purely from K-theoretic data. The proof hinges on resolving a long-standing question from 1988 posed by John Fogarty about whether differential forms on finite quotient spaces behave in the expected way, which the authors answer affirmatively.
On the other hand, the paper also shows that K-regularity cannot always detect bad geometry. The authors construct explicit projective spaces with finite quotient singularities that are K-m-regular for every m, yet fail to be local complete intersections, a basic geometric niceness condition weaker than smoothness. This means that in the projective setting, K-regularity is insufficient to force the space to be smooth or even particularly well-behaved. Together, the positive and negative results draw a clearer map of exactly when and where K-theory can serve as a reliable probe of geometric regularity.