Research

String theory admits an enormous number of consistent vacua, and ours is one of them. It is Borges's Library of Babel with universes for volumes, pick a book off the shelf, why did you pick this one?

Vacuum selection

Within the KKLT (Kachru, Kallosh, Linde, Trivedi) class of flux compactifications, a universe decays from one vacuum to the next by Coleman–De Luccia tunnelling. I found that the metastable vacua carrying an approximate R-symmetry are the most stable places for that chain to stop, that many of them come with parametric control over supersymmetry breaking, and that the same symmetry bears on the flatness problem and CP violation. The estimate ties the thin-wall instanton on the gravity side to the arithmetic of fluxes threading cycles of the internal geometry on the other. The papers, with Michael Dine and Guido Festuccia, are Metastable Domains of the Landscape, Symmetric Points in the Landscape as Cosmological Attractors and The Fate of Nearly Supersymmetric Vacua.

A matrix model of de Sitter space

Quantum mechanics requires probabilities to sum to one, which is awkward in a cosmology with a horizon. With Tom Banks and Bartomeu Fiol I proposed a fermionic matrix theory that uses the Hopf map to produce spacetime holographically, with room for particle states and for large deformations such as black holes. We conjectured that the large but finite size of the matrices is tied to cosmological supersymmetry breaking, a connection that Susskind's later work develops.

Supergravity

Supergravity is the low-energy spacetime action of a supersymmetric string theory. I built model laboratories in it for testing how supersymmetry breaks in different string theories, at the interface with low-energy particle phenomenology: moduli stabilisation in type IIA with badly broken supersymmetry, and moduli decays and the gravitino problem.

What carried over

If one can write down an action, one can model the dynamics, and that holds for a generative network as it does for a gauge theory.