Talks

Topological strings are well understood in perturbation theory, but their non-perturbative structure has been the subject of much work and speculation. A useful approach to this problem is to unveil the non-perturbative sectors of the theory by looking at the large order behavior of the perturbative series. This approach can be formulated in a mathematically rigorous way by using the theory of resurgence. In this talk I review the basic ideas behind this approach and I show that it can be successfully applied to topological string theory on arbitrary Calabi-Yau threefolds. Thsee methods lead, not only to explicit non-perturbative topological string amplitudes, but also to a conjecture relating the “invariants” of the theory of resurgence (also known as Stokes constants) to BPS invariants. Physically, this conjecture implies that the large order behavior of the topological string perturbative series contains information about the spectrum of stable D-brane sectors.

---

Zoom link

Besides providing a possible explanation to the strong CP problem and dark matter, the QCD axion possesses a rich cosmology. For example, if PQ breaking occurs after inflation, then axion cosmic strings form. Near the QCD phase transition, every axion string become attached to a domain wall which pull on the strings and cause the string-wall network to decay. While every string becomes attached to a domain wall, it is possible, though rare, to form an enclosed domain wall that is not attached to any axion string. These enclosed domain walls collapse under their own self tension, compressing a large amount of energy into a small volume and thereby potentially forming a primordial black hole. In this talk, I will discuss the abundance of enclosed domain walls, their dynamics of collapse, the efficiency of black hole formation, and their relic abundance. For sufficiently large axion decay constants, there may be an observable gravitational lensing signal at future lensing telescopes, especially in models with axion-like-particles.

---

Zoom link

Both the social and natural world are replete with complex structure that often has a probabilistic interpretation. In the former, we may seek to model, for example, the distribution of natural images or language, for which there are copious amounts of real world data. In the latter, we are given the probabilistic rule describing a physical process, but no procedure for generating samples under it necessary to perform simulation. In this talk, I will discuss a generative modeling paradigm based on maps between probability distributions that is applicable to both of these circumstances. I will describe a means for learning these maps in the context of problems in statistical physics, how to impose symmetries on them to facilitate learning, and how to use the resultant generative models in a statistically unbiased fashion. I will then describe a paradigm that unifies flow-based and diffusion based generative models by recasting generative modeling as a problem of regression. I will demonstrate the efficacy of doing this in computer vision problems and end with some future challenges and applications.

---

Zoom link