Scientific Series

I will present a brief introduction to non-Lorentzian geometries, an important example of such geometries being Newton-Cartan geometry and its torsionful generalization, which is the natural geometry to which non-relativistic field theories couple to. The talk will subsequently review how such geometries have in recent years appeared in gravity, string theory and holography. In particular, torsional Newton-Cartan geometry has been shown to appear as the boundary geometry for Lifshitz spacetimes. Furthermore, dynamical Newton-Cartan geometry is related to Horava-Lifsthiz gravity theories and appears in novel Chern-Simons theories of gravity in three dimensions. The latter can be obtained from a well-defined limit of the AdS3/CFT2 correspondence. Finally, I will briefly comment on how Newton-Cartan geometry appears in non-relativistic string theory.

The Bondi mass loss formula has been central in the context of early research on gravitational waves. We show how it can be understood as a particular case of BMS current algebra and discuss the associated central extension. We then move down to three dimensions where a more complete picture emerges.

Suppose the eigenvalue distributions of two matrices $M_1$ and $M_2$ are known. What is the eigenvalue distribution of the sum $M_1+M_2$? This problem has a rich pure mathematics history dating back to H. Weyl (1912) with many applications in various fields. Free probability theory (FPT) answers this question under certain conditions, which often involves some degree of randomness (disorder). We will describe FPT and show examples of its powers for the qualitative understanding (often approximations) of physical quantities such as density of states, and gapped vs. gapless phases of some Floquet systems. These physical quantities are often hard to compute exactly. Nevertheless, using FPT and other ideas from random matrix theory excellent approximations can be obtained. Besides the applications presented, we believe the techniques will find new applications in new contexts.

There has been a dissatisfaction with the postulates of quantum mechanics essentially since the moment that those postulates were first written down. Over the years since there have therefore been many attempts (some successful and some less so) to reconstruct quantum theory from various sets of postulates. The aim being to gain a deeper understanding of the theory by providing a conceptually clear underpinning from which the standard formalism can be derived. In this talk I present recent work with Carlo Maria Scandolo and Bob Coecke (arXiv:1802.00367) in which we reconstruct quantum theory from purely diagrammatic postulates within the process theory framework, showing that the conceptual bare-bones of quantum theory concerns the manner in which systems and processes compose.

Generically, a small amount of matter introduced to anti-de Sitter spacetime leads to formation of a black hole; however, the high degree of symmetry of AdS means that some initial distributions of matter (possibly also technically generic) oscillate indefinitely. Whether a given initial profile leads to a horizon at arbitrarily small amplitudes is of great interest for a number of reasons, not least because horizon formation corresponds holographically to thermalization in CFT.  We will present an overview of approaches the question and show a phase diagram of the stability behavior of AdS in the presence of a scalar field, including an analysis of nonperturbative phases at the border of perturbative stability and instability.

The next frontiers in cosmic microwave background (CMB) science include a detailed mapping of the CMB polarization anisotropy, with goals of detecting the inflationary B-mode signal and reconstructing high-fidelity maps of the matter distribution via CMB lensing, as well as a first detection of CMB spectral distortions.  At this level of precision (~nK), Galactic and extragalactic foregrounds may be the ultimate limiting factor in deriving cosmological constraints.  I will discuss biases due to foregrounds in CMB lensing measurements, including the first calculation of the lensing bias due to the kinematic Sunyaev-Zel’dovich effect, as well as recent progress in developing novel foreground-free CMB lensing estimators.  I will then present methods to extend CMB foreground parameterizations in a systematic, flexible way, with applications to both polarization and spectral distortion measurements.  Using this framework, I will discuss spectral distortion forecasts for CMB spectrometer mission concepts, showing that high-significance measurements of the Compton-y and relativistic thermal Sunyaev-Zel’dovich signals can be expected, as well as a potential detection of the primordial mu-type distortion due to Silk damping of small-scale acoustic modes.

Over the last decade, there have been enormous gains in machine learning technology primarily driven by neural networks. A major reason neural networks have outperformed older techniques is that the cost of optimizing them scales well with the size of the training dataset. But neural networks have the drawback that they are not very well understood theoretically.

Recent work by several groups has explored an alternative approach to creating machine learning model functions based on tensor networks, which are a technique for parameterizing many-body quantum wavefunctions. The cost of training tensor network models scales similarly to the cost of training neural networks. In addition, their relatively simple, linear structure has provided good theoretical understanding of their properties, and underpins many powerful techniques to optimize and manipulate them. 

After introducing tensor network machine learning models, I will discuss some of the techniques to optimize them and results for supervised and generative machine learning tasks. These techniques can automatically adapt the number of parameters and suggest interesting interpretations and extensions for 'deep' tensor network variants. I will conclude by discussing a recent tensor network based proposal to formulate hybrid quantum-classical algorithms for machine learning with quantum computers.

In this talk we will discuss three (intimately related) examples of 4-manifold invariants based on higher structures:
   - VOA[M4] from transgression of EFTs;
   - SW and Donaldson invariants as chiral algebra correlators;
   - Massey triple products from trisections.
These topics are based, respectively, on work with A.Gadde, P.Putrov (and work to appear with B.Feigin); recent paper with M.Dedushenko and P.Putrov; and a solo paper of the speaker.

Skew Howe duality is based on a very simple observation: the set of n by m matrices has commuting action of SL_n and SL_m.  We can use this observation to study morphisms of GL_m representations using GL_n.  This perspective has proven very useful in recent years for studying quantum knot invariants and their categorifications.  I will survey work in this direction from the last 10 years, including more recent developments concerning annular skew Howe duality and annular knot invariants.

In this talk, intended for a broader audience, I promise to use techniques only at the level of university calculus. While staying at this level, our goal will be to learn conceptual lessons for categorification of quantum group invariants of knots and 3-manifolds, also known as the Witten-Reshetikhin-Turaev (or WRT) invariants. In particular, we will introduce new q-series invariants of 3-manifolds that have integer powers and integer coefficients and, if time permits, discuss their various constructions and properties. The talk is based on several papers with D. Pei and/or P. Putrov, C. Vafa, as well as ongoing work with C. Manolescu.