The paper studies a special class of polynomial maps called graded Keller maps, which are related to a famous unsolved problem in mathematics called the Jacobian conjecture. This conjecture asks whether certain polynomial maps that look invertible based on a simple algebraic test are actually invertible. The paper organizes these maps by a numerical signature called a weight vector, and asks which signatures actually produce interesting counterexample candidates and how richly they are populated. The authors find that the answer depends heavily on whether the weights have a hyperbolic character, meaning some are positive and some negative.
To explore this, the authors build two explicit families of such maps in three variables. The first family, called cyclic cones, produces maps at a specific weight signature for infinitely many complexity levels, including examples defined over number fields involving square roots, and one particularly intricate example with 33 sheets defined over the ordinary rational numbers. The second family works at a different class of weight signatures and produces maps whose Jacobian determinant is a prime number rather than one, which is itself unusual. For both families, the authors analyze the symmetry of the maps using a concept from topology called geometric monodromy, finding that the symmetry group is as large as possible in almost all cases. This means that any algebraically simple, or solvable, behavior is confined to low complexity.
The most striking result concerns maps of a specific moderate complexity, where the authors show that the natural parameter space for these maps forms a smooth algebraic curve of a classical type called a plane cubic. Because this curve has infinitely many points and each point gives a genuinely distinct map under the natural notion of equivalence, the collection of such maps is not just a finite list of isolated examples but a continuously varying family with true moduli. This rules out the possibility that the subject is essentially finite or discrete, and places the study of graded Keller maps firmly within the realm of algebraic geometry with genuine geometric richness.