Bulletin of the American Physical Society
51st Annual Meeting of the APS Division of Atomic, Molecular and Optical Physics
Volume 65, Number 4
Monday–Friday, June 1–5, 2020; Portland, Oregon
Session P08: FOCUS: Machine Learning and Other New Ideas for Cold AtomsFocus Live
|
Hide Abstracts |
Sponsoring Units: DCOMP GDS Chair: Ehsan Khatami, San Jose State University Room: Portland 255 |
Thursday, June 4, 2020 2:00PM - 2:30PM Live |
P08.00001: String patterns and machine learning: probing the Fermi-Hubbard model with cold atoms Invited Speaker: Annabelle Bohrdt Cold atoms provide a versatile platform to study quantum many-body physics from a new perspective. They enable insights into systems that are challenging to describe theoretically and at the same time difficult to realize with a comparable amount of isolation, control, and tunability in solid state systems. The possibilities of cold atom experiments and quantum gas microscopy in particular pose a new opportunity and challenge for theorists to study novel observables that are now accessible. Recently, we have seen dramatic progress in the quantum simulation of the Fermi-Hubbard model, which in 2D is believed to capture essential features of high-temperature cuprate superconductors. In this talk I will present our recent work on the doped Fermi-Hubbard model. We study two-point spin correlation functions as well as less conventional higher order correlations relative to a dopant, which are not directly accessible in solid state experiments. Motivated by an intuitive picture of the motion of the dopant in the spin background, we search for string patterns in single snapshots taken with a quantum gas microscope. For an unbiased comparison of theories and experiment, we apply machine learning to classify experimental data at finite doping into different theoretical categories in order to determine which theory describes the system best on the microscopic level. [Preview Abstract] |
Thursday, June 4, 2020 2:30PM - 2:42PM Live |
P08.00002: Visualizing Correlations in the 2D Fermi-Hubbard Model with AI Ehsan Khatami, Elmer Guardado-Sanchez, Benjamin M. Spar, Juan Felipe Carrasquilla, Waseem S. Bakr, Richard T. Scalettar The physics of strongly correlated phases of matter is often described in terms of straightforward electronic patterns which are theoretically understood using Landau symmetry breaking theory. This has so far been the basis for studying correlations in the Fermi-Hubbard model realized with ultracold atoms. In this talk, we show that artificial intelligence (AI) can provide an unbiased alternative to this paradigm. Long and short range spin correlations spontaneously emerge in filters of a convolutional neural network trained on snapshots of single atomic species. In the less well-understood strange metallic phase of the model around 18\% doping, we find that a more complex network trained on snapshots of local moments produces an effective order parameter for the non-Fermi liquid phase. Our technique can be employed to characterize correlations unique to other phases with no obvious order parameters or signatures in projective measurements. [Preview Abstract] |
Thursday, June 4, 2020 2:42PM - 2:54PM Live |
P08.00003: Deep Convolutional Neural Networks for Quantum 1D Spin Chains Shah Saad Alam, Li Yang, Wenjun Hu, YiLong Ju, Han Pu, Ankit Patel Combining neural network architectures with quantum variational Monte Carlo methods has opened up a new method of studying quantum many body systems. Using deep learning to improve neural networks for quantum many body problems is a relatively new field of study. We discuss previous work in using a deep convolutional neural network for studying an SU(N) 1D spin chain, and our use of Importannce Sampling Gradient Optimization (ISGO) method to speed up the learning from the Variational Quantum Monte Carlo\footnote{ ``Deep Learning-Enhanced Variational Monte Carlo Method for Quantum Many-Body Physics'', Li Yang, Zhaoqi Leng, Guangyuan Yu, Ankit Patel, Wen-Jun Hu, Han Pu \url{https://arxiv.org/abs/1905.10730}}. We present our analysis of the neural network and the response of the networks layers to the particular symmetries of the SU(N) spin chain, as well as possible extensions of the neural network architecture. [Preview Abstract] |
Thursday, June 4, 2020 2:54PM - 3:06PM On Demand |
P08.00004: Machine learning aided study of a three-dimensional gas of SU($N$) fermions Entong Zhao, Jeongwon Lee, Chengdong He, Elnur Hajiyev, Zejian Ren, Toby T.H. Mak, Gyu-Boong Jo Machine Learning (ML) techniques have introduced a new scientific approach in quantum matter research. Recently, various quantum systems have been efficiently investigated based on ML-enabled analysis with synthetic data sets created from underlying theories, however, the direct application of ML-enabled analysis to the experimental data of quantum matter has remained elusive without fully utilizing useful attributes processed by the ML. In this talk, we report the ML-aided detection of SU($N$) fermions in which we apply ML techniques to classify nearly identical density profiles of ultracold Fermi gases based on their nuclear spin configurations. This ML-aided detection allows us to measure the spin multiplicity unattainable by any other method in a single shot. By testing and re-training of the neural networks (NN) with properly manipulated experimental images, we analyze how specific attributes of the density profile affect the classification accuracies of the NN. Our work demonstrates a new approach to machine learning aided classification problems in quantum physics, creating a potential to reveal and identify hidden features in highly complex quantum matter images. [Preview Abstract] |
Thursday, June 4, 2020 3:06PM - 3:36PM On Demand |
P08.00005: Generative modeling of quantum simulators Invited Speaker: Giacomo Torlai The recent advances in qubit manufacturing and coherent control of synthetic quantum matter are leading to a new generation of intermediate-scale quantum hardware, with promising progress towards simulation of quantum matter and materials. In order to enhance the capabilities of this class of quantum devices, some of the more arduous experimental tasks can be off-loaded to classical algorithms running on conventional computers. In this talk, I will present a framework based on industry-standard machine learning algorithms to perform approximate quantum state reconstruction from qubit measurement data. The resulting neural-network wavefunctions can be deployed to perform measurements of observables that may not be directly accessible in the original experimental platform, or that may entail a substantial experimental overhead in terms of both quantum resources and classical post-processing. I will demonstrate this approach for extracting the entanglement entropy from experimental cold-atom data, and for variational quantum chemistry simulations using superconducting qubits. [Preview Abstract] |
Thursday, June 4, 2020 3:36PM - 3:48PM |
P08.00006: Towards Machine Learning-Enhanced Laser Cooling and Trapping Shangjie Guo, Justyna P. Zwolak, I. B. Spielman Since its invention in the early 1970s, laser manipulation of atoms has broken records and became a routine technology. Today, researchers combine heuristics and out-of-loop optimization to inform the parameters in otherwise scripted experimental control sequences. However, as the configurations of both the apparatus (magneto-optical trap, MOT) and protocol (laser pulse sequences, magnetic fields, etc.) are numerous, the limits of laser cooling remain unknown. We will use machine learning (ML) techniques to optimize the MOT loading process and the following sub-Doppler laser cooling. In particular, we are developing a realistic 3D simulator of the experimental system to generate data for training an ML algorithm and to use as an off-line test bed. In this talk, I will discuss the design and preliminary performance of the simulator in comparison with experimental data for validation. [Preview Abstract] |
Follow Us |
Engage
Become an APS Member |
My APS
Renew Membership |
Information for |
About APSThe American Physical Society (APS) is a non-profit membership organization working to advance the knowledge of physics. |
© 2024 American Physical Society
| All rights reserved | Terms of Use
| Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 20740-3844
(301) 209-3200
Editorial Office
100 Motor Pkwy, Suite 110, Hauppauge, NY 11788
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700