My research focuses on the topology of stratified spaces. In particular I use
methods from algebraic topology to construct new algebraic invariants associ- ated to certain classes of stratified spaces, including many examples arising in theoretical physics and algebraic geometry. Theoretical physics provides a plethora of interesting examples, often in the form of moduli spaces. Such spaces are often smooth, non-compact manifolds. A common way to study a non-compact manifold is to study a particular compact- ification of it. This compactification frequently has the structure of a stratified space, meaning that it looks mostly like the smooth non-compact manifold, but instead of ’going of to infinity’ it instead degenerates into ’singularities.’ For example replace an infinite cylinder with the suspension of a circle, which looks like two hollow cones glued together along the common base circle. This has the effect of replacing the non-compact cylinder with a compact space. The trade off is that we get rid of non-compactness, but we instead have our space de- generating into singularities at the cone points. The reason this is useful comes from the fact that in the most simple cases topologically defined cohomology theories behave the same at cone points as analytically defined cohomology the- ories behave near infinity. This can be seen when comparing the intersection cohomology of a point with the L2 cohomology of a conical metric at infinity. To study a moduli space a physicists would view it as a non-compact com- plete manifold endowed with metric that behaves nicely near infinity. One can then produce various types of analytic invariants associated to the moduli space. These invariants usually represent, or measure something physical, such as D- branes, or configuration states of particles. It is often desired to know that these analytic ’invariants’ are actually invariant under certain changes. One of the goal of mathematicians is to describe these invariants in a topologically invariant way, and to therefore ensure that they have the desired properties. Many famous theorems do just this. The Hopf index theorem relates the index of a vector field on a smooth manifold to the Euler number of the manifold. The Atiya-Patodi-Singer index theorem relates the analytic index of an op- erator on a certain type of manifold to a topological index associated to the manifold. Cheeger’s Hodge theorem relates L2 -cohomology on a non-compact manifold endowed with a particular metric to the intersection homology defined by Goreskey and MacPherson. There are many, many more such theorem. Much of my work thus far has focused constructing topological spaces and cohomology theories associated to certain classes stratified spaces. In joint work with Markus Banagl, we have given the construction of an intersection space associated to certain types of stratified pseudomanifolds. This definition gen- eralizes the previous definition of Banagl, and allows for a much wider class of spaces to be analyzed. The goal of an intersection space is to produce a 1 new type of cohomology theory which satisfies Poincaré duality, but has slightly better properties than Goreskey and MacPherson’s intersection homology. The intersection space homology HI will, for example, have an internal cup prod- uct, while intersection homology IH does not in general have this. After we construct this intersections space we define a duality obstruction, and prove that the space satisfies Poincaré duality if and only if the duality obstruction vanishes. We then use this duality to prove certain vanishing statement about how the intersection homology signature of a Witt spaces will vanish when it satisfies certain conditions. With my colleague, Eugenie Hunsicker, we are trying to use some of the ideas and constructions from my work with Markus Banagl to characterize cer- tain L2 and weighted L2 cohomology theories topologically. These L2 -theories arise naturally as cohomology theories associated to certain types of moduli spaces of states in theoretical physics. Such spaces are non-compact manifold endowed with metrics which behave in a certain way near infinity. These metrics fall into various categories depending on their behavior: Asymptotically locally Euclidean (ALE), and asymptotically locally flat (ALF), among others. In each of these cases, as one approaches infinity, the space begins to looks like an open interval crossed with the total space of a fiber bundle π : E → B. In a recent paper Banagl and Hunsicker characterize the ALE case topologically using Ba- nagl’s intersection spaces. However the ALF case is still not well understood because as one approaches infinity the metric swaps the role of the base and fiber in the fiber bundle π. This is not too much of a problem if the bundle is trivial E ∼= F × B, but otherwise it introduces complications. Our main goal to define a meaningful cohomology theory associated to a certain compactifications of such an ALF space, and relate it to either the L2 -cohomology or the weighted L2 -cohomology of the ALF metric. A small side project that I have been working on recently is attempting to use the change of perversity map in intersection spaces to build a topological space which detects a type of L2 -cohomology associated to the stratified space from which the intersection space was constructed. Much of the motivation from this came from a careful examination of the proof of the main theorem in the above mentioned Banagl, Hunsicker paper. Previously in my thesis work I show how one could use the Whitney-de Rham complex, defined by my adviser Markus Plfaum, to determine the real-homotopy type of a semi-analytic set. I use a rectalinearization process pioneered by Bierstone and Milman in the late 1980s. This process allowed me to reduce the question to studying only semi-analytic sets which have singularities that are locally a normal crossing. As such spaces are locally isomorphic to a union of quadrants in some Euclidean space, I was able to use a smooth radial homotopy to derive the desired result. This method significantly simplifies the proof of a similar statement by Pflaum and Brasselet. As an application of this work we show how the Whitney-de Rham complex can be theoretically used to compute the cohomology of the free loop space of a semi-analytic set. Recently, Luca Prelli has used the language of constructable sheaves to claim a similar result for sub-analytic sets. Though the paper does not work out all of the details, Luca has in personal communications with me, given me a satisfactory explanation of the methods, and I believe the claim to be true.