Search

Ultracold atomic gases have revolutionized the study of non-equilibrium dynamics in quantum many-body systems. Many counterintuitive non-equilibrium effects have been observed, such as suppressed thermalization in a one-dimensional (1D) gas, the formation of repulsive self-bound dimers, and identical behaviors for attractive and repulsive interactions. Here, we observe the expansion of a bundle of ultracold 1D Bose gases in a flat-bottomed optical lattice potential. By combining in situ measurements with photoassociation, we follow the spatial dynamics of singly, doubly, and triply occupied lattice sites. The system sheds interaction energy by dissolving some doublons and triplons. Some singlons quantum distill out of the doublon center, while others remain confined. Our Gutzwiller mean-field model captures these experimental features in a physically clear way. These experiments might be used to study thermalization in systems with particle losses or the evolution of quantum entanglement, or if applied to fermions, to prepare very low entropy states., Comment: 22 pages, 8 figures

We theoretically study how excitations due to spontaneous emission and trap fluctuations combine with elastic collisions to change the momentum distribution of a trapped non-degenerate one-dimensional (1D) Bose gas. Using calculated collisional relaxation rates, we first present a semi-analytical model for the momentum distribution evolution to get insight into the main processes responsible for the system dynamics. We then present a Monte-Carlo simulation that includes features that cannot be handled analytically, and compare its results to experimental data. These calculations provide a baseline for how integrable 1D Bose gases evolve due to heating processes in the absence of diffractive collisions that might thermalize the gases.

We demonstrate Babinet's principle by the absorption of high intensity light from dense clouds of ultracold atoms. Images of the diffracted light are directly related to the spatial distribution of atoms. The advantages of employing Babinet's principle as an imaging technique are that it is easy to implement and the detected signal is large. We discuss the regimes of applicability of this technique as well as its limitations., Comment: 4 pages, 2 figures

We describe theoretical models for the recent experimental observation of Macroscopic Quantum Self-Trapping (MQST) in the transverse dynamics of an ultracold bosonic gas in a 2D lattice. The pure mean-field model based on the solution of coupled nonlinear equations fails to reproduce the experimental observations. It greatly overestimates the initial expansion rates at short times and predicts a slower expansion rate of the cloud at longer times. It also predicts the formation of a hole surrounded by a steep square fort-like barrier which was not observed in the experiment. An improved theoretical description based on a simplified Truncated Wigner Approximation (TWA), which adds phase and number fluctuations in the initial conditions, pushes the theoretical results closer to the experimental observations but fails to quantitatively reproduce them. An explanation of the delayed expansion as a consequence of a new type of self-trapping mechanism, where quantum correlations suppress tunneling even when there are no density gradients, is discussed and supported by numerical time-dependent Density Matrix Renormalization Group (t-DMRG) calculations performed in a simplified two coupled tubes set-up.

Structured prediction tasks pose a fundamental trade-off between the need for model complexity to increase predictive power and the limited computational resources for inference in the exponentially-sized output spaces such models require. We formulate and develop the Structured Prediction Cascade architecture: a sequence of increasingly complex models that progressively filter the space of possible outputs. The key principle of our approach is that each model in the cascade is optimized to accurately filter and refine the structured output state space of the next model, speeding up both learning and inference in the next layer of the cascade. We learn cascades by optimizing a novel convex loss function that controls the trade-off between the filtering efficiency and the accuracy of the cascade, and provide generalization bounds for both accuracy and efficiency. We also extend our approach to intractable models using tree-decomposition ensembles, and provide algorithms and theory for this setting. We evaluate our approach on several large-scale problems, achieving state-of-the-art performance in handwriting recognition and human pose recognition. We find that structured prediction cascades allow tremendous speedups and the use of previously intractable features and models in both settings., Comment: 32 pages, in submission

We study the transverse expansion of arrays of ultracold $^{87}$Rb atoms weakly confined in tubes created by a 2D optical lattice, and observe that transverse expansion is delayed because of mutual atom interactions. A mean-field model of a coupled array shows that atoms become localized within a roughly square fort-like self-trapping barrier with time-evolving edges. But the observed dynamics is poorly described by the mean-field model. Theoretical introduction of random phase fluctuations among tubes improves the agreement with experiment, but does not correctly predict the density at which the atoms start to expand with larger lattice depths. Our results suggest a new type of self-trapping, where quantum correlations suppress tunneling even when there are no density gradients.

We demonstrate 3D microwave projection sideband cooling of trapped, neutral atoms. The technique employs state-dependent potentials that enable microwave photons to drive vibration-number reducing transitions. The particular cooling sequence we employ uses minimal spontaneous emission, and works even for relatively weakly bound atoms. We cool 76% of atoms to their 3D vibrational ground states in a site-resolvable 3D optical lattice., Comment: 5 pages, 4 figures, Supplemental Material included. To appear in Physical Review Letters

We numerically calculate the reliability with which one can optically determine the presence or absence of an individual scatterer in a randomly occupied 3D array of well-localized, coherently radiating scatterers. The reliability depends in a complicated way on the ratio of lattice spacing to wavelength and the numerical aperture of the imaging system. The behavior can be qualitatively understood by considering the dependence of Bragg scattering modes on lattice spacing. These results are of interest for atomic implementations of quantum information processing.