Publications

Machine learning algorithms relying on deep neural networks recently allowed a great leap forward in artificial intelligence. Despite the popularity of their applications, the efficiency of these algorithms remains largely unexplained from a theoretical point of view. The mathematical description of learning problems involves very large collections of interacting random variables, difficult to handle analytically as well as numerically. This complexity is precisely the object of study of statistical physics. Its mission, originally pointed toward natural systems, is to understand how macroscopic behaviors arise from microscopic laws. Mean-field methods are one type of approximation strategy developed in this view. We review a selection of classical mean-field methods and recent progress relevant for inference in neural networks. In particular, we remind the principles of derivations of high-temperature expansions, the replica method and message passing algorithms, highlighting their equivalences and complementarities. We also provide references for past and current directions of research on neural networks relying on mean-field methods.

Ta
2
NiSe
5
is one of the most promising materials for hosting an excitonic insulator ground state. While a number of experimental observations have been interpreted in this way, the precise nature of the symmetry breaking occurring in
Ta
2
NiSe
5
, the electronic order parameter, and a realistic microscopic description of the transition mechanism are, however, missing. By a symmetry analysis based on first-principles calculations, we uncover the discrete lattice symmetries which are broken at the transition. We identify a purely electronic order parameter of excitonic nature that breaks these discrete crystal symmetries and contributes to the experimentally observed lattice distortion from an orthorombic to a monoclinic phase. Our results provide a theoretical framework to understand and analyze the excitonic transition in Ta_2NiSe_5 and settle the fundamental questions about symmetry breaking governing the spontaneous formation of excitonic insulating phases in solid-state materials.

We show how a large family of interacting nonequilibrium phases of matter can arise from the presence of multiple time-translation symmetries, which occur by quasiperiodically driving an isolated, quantum many-body system with two or more incommensurate frequencies. These phases are fundamentally different from those realizable in time-independent or periodically driven (Floquet) settings. Focusing on high-frequency drives with smooth time dependence, we rigorously establish general conditions for which these phases are stable in a parametrically long-lived “preheating” regime. We develop a formalism to analyze the effect of the multiple time-translation symmetries on the dynamics of the system, which we use to classify and construct explicit examples of the emergent phases. In particular, we discuss time quasicrystals which spontaneously break the time-translation symmetries, as well as time-translation symmetry-protected topological phases.

A notion of quantum natural evolution strategies is introduced, which provides a geometric synthesis of a number of known quantum/classical algorithms for performing classical black-box optimization. Recent work of Gomes et al. [2019] on combinatorial optimization using neural quantum states is pedagogically reviewed in this context, emphasizing the connection with natural evolution strategies. The algorithmic framework is illustrated for approximate combinatorial optimization problems, and a systematic strategy is found for improving the approximation ratios. In particular it is found that natural evolution strategies can achieve state-of-art approximation ratios for Max-Cut, at the expense of increased computation time.

We present an extensive ALMA spectroscopic follow-up programme of the z=4.3 structure SPT2349−56, one of the most actively star-forming proto-cluster cores known, to identify additional members using their [C{\sc ii}] 158\,μm and \mbox{CO(4--3)} lines. In addition to robustly detecting the 14 previously published galaxies in this structure, we identify a further 15 associated galaxies at z=4.3, resolving 55±5\,per cent of the 870-μm flux density at 0.5\,arcsec resolution compared to 21\,arcsec single-dish data. These galaxies are distributed into a central core containing 23 galaxies extending out to 300\,kpc in diameter, and a northern extension, offset from the core by 400\,kpc, containing three galaxies. We discovered three additional galaxies in a red {\it Herschel\/}-SPIRE source 1.5\,Mpc from the main structure, suggesting the existence of many other sources at the same redshift as SPT2349−56 that are not yet detected in the limited coverage of our data. An analysis of the velocity distribution of the central galaxies indicates that this region may be virialized with a mass of (9±5)×1012\,M⊙, while the two offset galaxy groups are about 30 and 60\,per cent less massive and show significant velocity offsets from the central group. We calculate the [C{\sc ii}] and far-infrared number counts, and find evidence for a break in the [C{\sc ii}] luminosity function. We estimate the average SFR density within the region of SPT2349−56 containing single-dish emission (a proper diametre of 720\,kpc), assuming spherical symmetry, to be roughly 4×104\,M⊙\,yr−1\,Mpc−3; this may be an order of magnitude greater than the most extreme examples seen in simulations.

The Galactic bulge is a tumultuous dense region of space, packed with stars separated by far smaller distances than those in the Solar neighborhood. A quantification of the frequency and proximity of close stellar encounters in this environment dictates the exchange of material, disruption of planetary orbits, and threat of sterilizing energetic events. We present estimated encounter rates for stars in the Milky Way bulge found using a combination of numerical and analytical methods. By integrating the orbits of bulge stars with varying orbital energy and angular momentum to find their positions over time, we were able to estimate how many close stellar encounters the stars should experience as a function of orbit shape. We determined that ~80% of bulge stars have encounters within 1000 AU and that half of bulge stars will have >35 such encounters, both over a gigayear. Our work has interesting implications for the long-term survivability of planets in the Galactic bulge.

Many scientific disciplines rely on computational methods for data analysis, model generation, and prediction. Implementing these methods is often accomplished by researchers with domain expertise but without formal training in software engineering or computer science. This arrangement has led to underappreciation of sustainability and maintainability of scientific software tools developed in academic environments. Some software tools have avoided this fate, including the scientific library Rosetta. We use this software and its community as a case study to show how modern software development can be accomplished successfully, irrespective of subject area. Rosetta is one of the largest software suites for macromolecular modeling, with 3.1 million lines of code and many state-of-the-art applications. Since the mid 1990s, the software has been developed collaboratively by the RosettaCommons, a community of academics from over 60 institutions worldwide with diverse backgrounds including chemistry, biology, physiology, physics, engineering, mathematics, and computer science. Developing this software suite has provided us with more than two decades of experience in how to effectively develop advanced scientific software in a global community with hundreds of contributors. Here we illustrate the functioning of this development community by addressing technical aspects (like version control, testing, and maintenance), community-building strategies, diversity efforts, software dissemination, and user support. We demonstrate how modern computational research can thrive in a distributed collaborative community. The practices described here are independent of subject area and can be readily adopted by other software development communities.

Photo-induced edge states in low dimensional materials have attracted considerable attention due to the tunability of topological properties and dispersion. Specifically, graphene nanoribbons have been predicted to host chiral edge modes upon irradiation with circularly polarized light. Here, we present numerical calculations of time-resolved angle resolved photoemission spectroscopy (trARPES) and time-resolved resonant inelastic x-ray scattering (trRIXS) of a graphene nanoribbon. We characterize pump-probe spectroscopic signatures of photo-induced edge states, illustrate the origin of distinct spectral features that arise from Floquet topological edge modes, and investigate the roles of incoming photon energies and finite core-hole lifetime in RIXS. With momentum, energy, and time resolution, pump-probe spectroscopies can play an important role in understanding the behavior of photo-induced topological states of matter.

Galactic double neutron star systems have a tight mass distribution around ∼1.35M⊙, but the mass distribution of all known pulsars is broader. Here we reconstruct the Alsing, et al. (2018) bimodal mass distribution of pulsars observed in binary systems, incorporating data from observations of J0740+6620 which were not available at the time of that work. Because J0740+6620 is an outlier in the mass distribution with non-negligible uncertainty in its mass measurement, its mass receives a large correction from the population, becoming mJ0740+6620=2.03+0.10−0.08M⊙ (median and 68\% CI). Stochastic samples from our population model, including population-informed pulsar mass estimates, are available \href{https://github.com/farr/AlsingNSMassReplication}{at this https URL}.

Electron ratchets are non-equilibrium electronic devices that break inversion symmetry to produce currents from non-directional and random perturbations, without an applied net bias. They are characterized by strong parameter dependence, where small changes in operating conditions lead to large changes in the magnitude and even direction of the resulting current. This high sensitivity makes electron ratchets attractive research subjects, but leads to formidable challenges in their deeper study, and particularly to their useful application. This perspective reviews the progress that was made in the field starting from the first experimental electron ratchets in the late 1990s, and how the field spawned multiple designs with very different properties. We discuss the possible uses of electron ratchets in sensing and energy harvesting, and the specific issues encountered when idealized behavior meets complex reality. We promote an application-driven approach where complexity is not necessarily detrimental and argue that a system level perspective would be beneficial over reductionism. We highlight several promising research directions, which revolve around the intentional study of complex effects, and the modeling of realistic devices.