The paper tackles a problem in algebraic geometry and singularity theory, which studies the behavior of geometric objects near points where they fail to be smooth. Specifically, it focuses on "complete intersections," which are geometric shapes defined by several polynomial equations simultaneously. The singularities of these shapes, meaning the points where they cross themselves or have sharp corners or cusps, can be measured and classified using various numerical invariants. Two important classes of such shapes are studied here: "semi-quasihomogeneous" complete intersections, which have a dominant term that controls the shape near the singular point, and "Khovanskii non-degenerate" complete intersections, where the singularity structure is governed by a combinatorial object called a Newton polyhedron.
The central quantity being computed is called the "minimal exponent," a refined numerical invariant associated to a singularity. It generalizes an older invariant called the log canonical threshold and carries more subtle information about how bad a singularity is. The minimal exponent is related to the spectrum of the singularity and has connections to questions in birational geometry, such as whether certain spaces have nice curvature properties. Computing it is generally quite hard, and explicit formulas are only known in special cases.
The authors derive explicit formulas for the minimal exponent in both of the geometric settings they consider. For semi-quasihomogeneous complete intersections, the answer can be expressed in terms of the weights and degrees that define the dominant part of the equations. For Khovanskii non-degenerate complete intersections, the answer is expressed in terms of the combinatorial data of the associated Newton polyhedra. These results extend earlier work done for the simpler case of a single polynomial equation defining a hypersurface, and they provide a concrete toolkit for understanding singularities in a much broader range of geometric situations.