Bulletin of the American Physical Society
APS March Meeting 2022
Volume 67, Number 3
Monday–Friday, March 14–18, 2022; Chicago
Session W13: Autonomous SystemsFocus Live Streamed
|
Hide Abstracts |
Sponsoring Units: GDS Chair: William Ratcliff, GDS Room: McCormick Place W-183A |
Thursday, March 17, 2022 3:00PM - 3:36PM |
W13.00001: Domain-Aware Gaussian Processes and High-Performance Mathematical Optimization for Optimal Autonomous Data Acquisition Invited Speaker: Marcus Noack Gaussian processes and Gaussian-related stochastic processes have been shown to be powerful tools for stochastic function approximation and autonomous control of data acquisition due to their robustness, analytical tractability, and natural inclusion of uncertainty quantification. In this talk, I want to present our work on a general, flexible, and powerful GP-driven framework for autonomous data acquisition. The focus of this work lies on making Gaussian (related) processes more flexible and domain aware, how the added flexibility and domain-awareness can be used for decision-making, and the computational and mathematical challenges that come with these advancements. In particular, I will focus on the challenge of hyperparameter optimization and how it impacts on-the-fly autonomous decision-making. |
Thursday, March 17, 2022 3:36PM - 4:12PM |
W13.00002: Quantum compiling by deep reinforcement learning Invited Speaker: Enrico Prati The architecture of circuital quantum computers requires computing layers devoted to compiling high-level quantum algorithms into lower-level circuits of quantum gates. The general problem of quantum compiling is to approximate any unitary transformation that describes the quantum computation, as a sequence of elements selected from a finite base of universal quantum gates. The existence of an approximating sequence of one qubit quantum gates is guaranteed by the Solovay-Kitaev theorem, which implies sub-optimal algorithms to establish it explicitly. Since a unitary transformation may require significantly different gate sequences, depending on the base considered, such a problem is of great complexity and does not admit an efficient approximating algorithm. Therefore, traditional approaches are time-consuming tasks, unsuitable to be employed during quantum computation. We exploit the deep reinforcement learning method as an alternative strategy, which has a significantly different trade-off between search time and exploitation time. Deep reinforcement learning allows creating single-qubit operations in real time, after an arbitrary long training period during which a strategy for creating sequences to approximate unitary operators is built. The deep reinforcement learning based compiling method allows for fast computation times, which could in principle be exploited for real-time quantum compiling. I review the results on one and two- qubits quantum logic gates and I show how deep reinforcement learning can be exploited to address optimizationo of problems involving qubits in quantum computing acrhitectures. |
Thursday, March 17, 2022 4:12PM - 4:48PM |
W13.00003: Dynamic Manipulation of Ferroelectric Structures via Automated Piezoresponse Force Microscopy Invited Speaker: Kyle Kelley Ferroelectric domain walls and other topological defects underpin a broad range of applications ranging from domain wall electronics to nonlinear optics. Moreover, exploring their functionalities is of interest to a broad spectrum of applications. However, the dynamic nature of these objects severely constrains piezoresponse force microscopy approaches to explore their functionality. Here, to overcome this challenge, we introduce two examples of automated experiments in piezoresponse force microscopy. First, a real-time piezoresponse force microscopy feedback approach is utilized to control the scanning probe bias during imaging allowing for modification of the domain structures triggered by detected states under the tip. Specifically, the complex trigger system (i.e., “FerroBot”) is used study metastable and frustrated domain-wall dynamics in a PbxSr(1-x)TiO3 system. We have demonstrated that frustrated phases with enhanced electromechanical response can be created using this technique. Further, we have identified a metastable mechanism of the domain-wall dynamics in this system, i.e., domain-wall bending can be separated from irreversible domain reconfiguration regimes. In conjunction, phase-field modeling was used to corroborate the observed mechanisms. Secondly, we introduce computer vision-based automated experiments to explore the ferroelectric switching behavior of elongated domains in BiFeO3 thin films. These studies highlight a new pathway towards the discovery and control of metastable states in ferroelectrics, and more generally paves way for automated systems for controlled modification of domain walls and defects to improve material properties. |
Thursday, March 17, 2022 4:48PM - 5:00PM |
W13.00004: Active Learning-Driven Automated Scanning Probe Microscopy Enables Discovery of Structure-Property Relationship Yongtao Liu, Kyle Kelley, Rama K Vasudevan, Hiroshi Funakubo, Susan E Trolier-Mckinstry, Maxim Ziatdinov, Sergei V. Kalinin The functionalities of topological and structural defects in ferroelectric materials including domain wall dynamics, conductivity of topological defects, and light-induced phenomena have been a source of much fascination in the condensed matter physics community. Until now, the search for interesting functionalities has been guided by auxiliary information from scanning probe microscopy (SPM) to identify potential objects of interest based on human intuition. Here, we developed a machine learning-based automated SPM workflow that actively discovers relationships between domain structure and functional responses (hysteresis loops, I-V curves, non-linearities). This automated workflow combines the power of machine learning methods to learn the correlative relationships between high dimensional data, and human-based physics insights encoded in the acquisition function. This approach demonstrates that the discovery path and sampling points of on-field and off-field hysteresis loops are largely different in a PbTiO3 thin film. The larger polarization mobility in the vicinity of 180o walls results in more significant hysteresis loop opening in on-field measurements, while off-field measurements detect only the slowly relaxing components due to stronger pinning at the ferroelastic walls. This approach can be adapted to apply to a broad range of imaging and spectroscopy methods, e.g., SPM, electron microscopy, optical microscopy, and chemical imaging. |
Thursday, March 17, 2022 5:00PM - 5:12PM |
W13.00005: Towards automating analysis of nonequilibrium X-ray Photon Correlation Spectroscopy with acquisition rate-limited time resolution Tatiana Konstantinova, Lutz Wiegart, Maksim Rakitin, Anthony M DeGennaro, Andi Barbour Non-equilibrium behavior is characteristic for many systems studied with X-ray Photon Correlation Spectroscopy (XPCS). Analysis of non-equilibrium XPCS data often involves manual iterative binning of quasi-equilibrium regions in an attempt to reach acceptable signal-to-noise ratios in the intensity-intensity correlation functions. Such approach often leads to loss of temporal resolution as well as accumulation of systematic errors in parameters quantifying the dynamics. In this talk, I present a new approach for automating the analysis of XPCS data that allows to preserve the temporal resolution set by the experimental acquisition rate. The analysis workflow involves use of denoising autoencoders, scaled to an arbitrary input size, algorithmic adjustment of dynamics parameters' bounds and estimation of analysis uncertainty beyond parameters' variances. |
Thursday, March 17, 2022 5:12PM - 5:24PM |
W13.00006: Adaptive tuning of the latent space of encoder-decoder convolutional neural networks for virtual 6D diagnostics of time-varying charged particle beams Alexander Scheinker Advanced and future high energy physics facilities, such as the FACET-II beam-driven plasma wakefield accelerator (PWFA), require an ability to generate and accelerate extremely short (few fs) high charge (few nC) high peak current (> 200 kA) electron bunches while maintaing minimal energy spread and precise control of current profiles which is crucial for the PWFA process. We present a novel method of adaptive machine learning (AML) which combines ML techniques such as convolutional neural networks (CNN) with model-independent feedback control theory for time-varying systems. Our encoder-decoder CNN maps high dimensional inputs (128x128 pixel images = 16348 dimensions) to representations in a general nonlinear 2-dimensional latent space, which is then used to generate a 1.2288E6 dimensional output which is all 15 of the unique 2D projections of a charged particle beam's 6D (x,y,z,px,py,E) phase space at 5 different locations in an accelerator (75 128x128 pixel images). Only a single one of the 75 generated projections, the (z,E) longitudinal phase space, is compared to a measurement which provides adaptive feedback acting directly on the low dimensional latent space of the network allowing us to estimate all projections of the beam's 6D phase space as the beam changes with time. |
Thursday, March 17, 2022 5:24PM - 5:36PM |
W13.00007: Autonomous anomaly detection in MeV ultrafast electron diffraction Mariana A Fazio, Salvador Sosa Guitron, Destry Monk, Junjie Li, Marcus Babzien, Mikhail Fedurin, Mark A Palmer, Sandra G Biedron, Manel Martínez-Ramón MeV ultrafast electron diffraction (MUED) is a pump-probe technique to measure dynamic material structure evolution. An ultrashort laser initiates a structure change which is probed by an ultrashort relativistic electron beam. Diffraction patterns are integrated over many shots to beat low signal-to-noise ratio. However, electron beam instabilities from shot to shot disturb the patterns and increase uncertainty. To enhance the accuracy of MUED, anomalous patterns should be detected and removed from datasets with thousands of patterns. |
Thursday, March 17, 2022 5:36PM - 5:48PM |
W13.00008: Reinforcement learning and neutron scattering William Ratcliff, Kate Meuse, Jessica Opsahl-Ong, Paul Kienzle During this talk, I will discuss our recent progress with applying reinforcement learning to neutron scattering. Examples include single crystal diffraction, measurements of the order parameter, and measurement of spin-wave excitations. Our results are currently on simulated data and show that it is possible to use reinforcement learning to dramatically reduce the number of measurements required to obtain parameters from experiments. I will also discuss the advantages of incorporating physics into models |
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