Bulletin of the American Physical Society
APS March Meeting 2021
Volume 66, Number 1
Monday–Friday, March 15–19, 2021; Virtual; Time Zone: Central Daylight Time, USA
Session P33: Quantum Control in the Presence of Noise and DecoherenceFocus Live

Hide Abstracts 
Sponsoring Units: DQI Chair: Christopher Eichler, Princeton University 
Wednesday, March 17, 2021 3:00PM  3:12PM Live 
P33.00001: Deep reinforcement learning for quantum Hamiltonian engineering Pai Peng, Xiaoyang Huang, Chao Yin, Chandrasekhar Ramanathan, Paola Cappellaro Engineering desired Hamiltonians in quantum manybody systems is essential for applications such as quantum simulation, computation and sensing. Conventional Hamiltonian engineering sequences are designed using human intuition based on perturbation theory, which may not be optimal and is unable to accommodate complex experimental imperfections. Here we search for Hamiltonian engineering sequences using deep reinforcement learning (DRL) and experimentally demonstrate that they outperform celebrated sequences in a solidstate nuclear magnetic resonance quantum simulator. We aim at decoupling stronglyinteracting spin1/2 systems and consider different experimental imperfections. The robustness of the DRLfound sequences is verified both in simulations and experiments. More interestingly, many of the DRL sequences exhibit a common pattern unknown before. By restricting the searching space to such patterns, we ultimately design sequences that are robust against dominant imperfections in our experiments. We not only demonstrate a general method for Hamiltonian engineering, but also highlight the importance of both blackbox artificial intelligence and understanding of physical system in order to realize experimentally viable applications. 
Wednesday, March 17, 2021 3:12PM  3:24PM Live 
P33.00002: Reinforcement Learningbased control of a cavity system with non linear measurement Riccardo Porotti, Florian Marquardt, Antoine Essig, Audrey Bienfait, Benjamin Huard Quantum states of the radiation field inside cavities have emerged as a promising new platform for quantum information processing. One major challenge is to find efficient ways to prepare nonclassical quantum states. Recently, deep reinforcement learning (DRL) has been introduced into quantum physics applications. We show that indeed DRL techniques can discover such strategies from scratch, by employing feedback based on nonlinear measurements, and create nonclassical states even in the absence of nonlinear controls. 
Wednesday, March 17, 2021 3:24PM  3:36PM Live 
P33.00003: Classical Reinforcement Learning for Experimental Quantum Error Correction Volodymyr Sivak, Henry Liu, Alec Eickbusch, Baptiste Royer, Ioannis Tsioutsios, Michel Devoret Quantum feedback control framed as a partially observable Markov decision process poses a significant challenge for reinforcement learning methods due to the minimalistic observability of quantum states through projective measurements. To alleviate this, all applications of reinforcement learning to quantum control to date rely on the simulation of the system which gives the learning agent direct access to the wavefunction. In contrast, using modelfree reinforcement learning, we solve a challenging task of learning the parameters of quantum error correction protocol for bosonic gridstate logical encoding directly from ancilla qubit binary measurement outcomes. Such approach, applied to a quantum system whose wavefunction is not directly accessible in a real experiment, will completely eliminate the model bias and allow to learn quantum control policies tailored to the specific analog error channels present in the experimental system. 
Wednesday, March 17, 2021 3:36PM  3:48PM Live 
P33.00004: Experimental Evaluation of Active Learning of a Two Qubit CrossResonance Hamiltonian Arkopal Dutt, Edwin Pednault, Chai Wu, Sarah Sheldon, John Smolin, Lev S Bishop, Isaac Chuang An important step in calibration and control is Hamiltonian learning, which involves learning the parameters given a Hamiltonian model and description of noise sources through queries to the quantum system. Standard techniques require $O(\epsilon^{2})$ queries to achieve learning error $\epsilon$ due to the standard quantum limit. To minimize the number of queries required and improve the scaling with $\epsilon$, we propose a Hamiltonian active learner based on Fisher information (“HALFI”). Each input query specifies the initial state, measurement operator and interaction time, and the resulting output is a single shot binary valued measurement. Seeded with data from an initial set of queries, HALFI optimizes subsequent queries. Performance of HALFI is evaluated on a twoqubit crossresonance gate on a 20qubit IBM Quantum device, using Qiskit Pulse to model readout noise, imperfect pulseshaping and decoherence. HALFI realizes a 27% reduction in resource requirements over an uniformly random approach, with an order of magnitude reduction over quantum process tomography. Moreover, by spacing out queries nonuniformly in time, HALFI can achieve learning error which scales inversely with the number of queries, meeting the Heisenberg bound. 
Wednesday, March 17, 2021 3:48PM  4:00PM Live 
P33.00005: Continuous measurements with feedback can improve singlephoton probability Anirudh Lanka, Todd Brun We propose a technique to improve the probability of singlephoton emission with an electrically pumped quantum dot in an optical microcavity, by continuously monitoring the dot’s energy state and using feedback to control whether to stop pumping. The goal is to boost the probability of singlephoton emission while bounding the probability of two or more photons. We model the system by a stochastic master equation that includes postmeasurement operations. Ideally, feedback will be based on the entire continuous measurement record, but in practice, it may be difficult to process such a protocol in real time. We show that even a simple thresholdbased feedback scheme using measurements at a single time can improve performance over deterministic (openloop) pumping. This technique is particularly useful for strong dotcavity coupling strength with lower rates of pumping, as can be the case for electrical pumping. It is also numerically tractable since we can perform ensemble averaging with a single master equation rather than averaging over a large number of quantum trajectories. 
Wednesday, March 17, 2021 4:00PM  4:12PM Live 
P33.00006: Optimal Quantum Control of TimeCorrelated Semiclassical Control Noise Colin Trout, Kevin Schultz, Greg Quiroz, Robert Barr, David Clader Precise control of quantum systems is required for most proposed quantum technologies. The major roadblock in precise quantum control of quantum systems is noise due to unwanted environmental interactions. Provably optimal control sequences for select systembath interactions exist while numerical optimization methods have traditionally been applied to mitigate errors in the control. In this work, we provide an analytical framework for constructing optimal control for systems under the influence of temporallycorrelated semiclassical control noise. This framework is built from Autoregressivemovingaverage models (ARMA) from time series analysis for representing signals with arbitrary noise spectra. With this framework, we have constructed provably optimal quantum control sequences for singleaxis quantum control under the influence of temporallycorrelated semiclassical control noise with select noise spectra. In this talk, I will discuss the framework, optimality of the solutions for select noise spectra, and extensions to signals with arbitrary spectra. 
Wednesday, March 17, 2021 4:12PM  4:24PM Live 
P33.00007: Using spectator qubit to Optimally mitigating qubit decoherence Hongting Song, Behnam Tonekaboni, Areeya Chantasri, Howard Wiseman To reduce or mitigate the decoherence caused by environmental noise on quantum systems, one first need to characterize the noise. This is the goal of "quantum noise spectroscopy" (QNS). QNS provide information about correlation functions which are average quantities over many noise realisations. Here we go one step further; equipped with the knowledge from QNS, we would like to track the effect of single noise realization. To do that, we introduce a spectator qubit which senses the same noise and we use it to monitor the noise and predict the evolution of the logical qubit. 
Wednesday, March 17, 2021 4:24PM  5:00PM Live 
P33.00008: Integration of spectator qubits into quantum computing architectures for adaptive hardware tuneup and noise characterization Invited Speaker: Riddhi Swaroop Gupta In this talk, we explore the implications of attempts to efficiently tuneup and calibrate mesoscale quantum computers on the optimal physical layout of the underlying quantum hardware. We introduce a previously proposed classical adaptive learning algorithm, Noise Mapping for Quantum Architectures (NMQA), and present several technical innovations that enable classical filtering of discrete projective measurements, useful for adaptively learning systemdynamics, noise properties or hardware performance variations in classically correlated measurement data. Using insights from sequential MonteCarlo approaches, convergence properties of NMQA are discussed. In the second part of the talk, we focus on the specific challenge of calibrating ("mapping") spatially inhomogeneous noise fields or calibration errors using spectator qubits dedicated to the problem of sensing. Drawing on optimal approximation theory to dictate sampling locations, we present optimal sensorqubit arrangements at the Padua points to reconstruct dephasing fields in 2D via Lagrange approximation methods. The performance of these Paduainspired techniques is compared to NMQA, using the same qubit arrangement on hardware. Our results show that Paduainspired techniques display optimal error scaling behavior relative to NMQA in ideal cases, and we probe the limits of these benefits as a function of measurement noise and spatial errors in qubit locations. Extensions to incorporate spatiotemporal dynamics are discussed. 
Wednesday, March 17, 2021 5:00PM  5:12PM Live 
P33.00009: Calibration of the CrossResonance Gate using ClosedLoop Optimal Control Brad Mitchell, Ravi K. Naik, Alexis Morvan, Akel Hashim, John Mark Kreikebaum, David Ivan Santiago, Irfan Siddiqi The crossresonance gate is an appealing entangling gate for scaling up superconducting quantum processors because of its modest hardware requirements and low error rates. However, reliably achieving coherencelimited fidelity for this gate is challenging due to the multitude of parameters to calibrate, especially in the presence of crosstalk and spectator qubits. Further, optimal parameters for the gate are sensitive to the detuning between the qubits and hence to fabrication variation. For calibrating crossresonance gates on a large quantum processor, a method that can efficiently optimize parameters across a set of qubits with a spread of qubit frequencies and crosstalk dynamics is desired. 
Wednesday, March 17, 2021 5:12PM  5:24PM Live 
P33.00010: Contextual Characterization of the CrossResonance Gate on a MultiQubit Superconducting Quantum Processor Ravi K. Naik, Brad Mitchell, Akel Hashim, John Mark Kreikebaum, David Ivan Santiago, Irfan Siddiqi The performance of algorithms on quantum processors is determined by the error rates of the single and twoqubit gates composing the algorithm. However, in multiqubit processors, the presence of both intrinsic and controlinduced crosstalk limits the efficacy of traditional gate error metrics for determining algorithmic performance. Determining the nature of and reducing the errors that occur on spectator qubits when gates are applied is essential. In this work, we use contextual benchmarks to characterize the performance of crossresonance gates in a superconducting quantum processor. We describe an iterative method where we determine specific errors on a multiqubit processor due to cycles containing crossresonance gates, and account for those errors using virtual gates and dynamical decoupling. Finally, we explore the utilization of the crossresonance effect for multiqubit gates. 
Wednesday, March 17, 2021 5:24PM  5:36PM Live 
P33.00011: Quantum optimal control for highfidelity arbitrary quantum logic on a superconducting qudit Xian Wu, Spencer Tomarken, N. Anders Petersson, Luis A. Martinez, Yaniv J Rosen, Kyle A Wendt, Konstantinos Kravvaris, Sofia Quaglioni, Jonathan L DuBois Quantum simulation is one of the most anticipated application of quantum computation, which often requires repeated applications of highfidelity arbitrary quantum logic on the quantum hardware. Arbitrary quantum logic is often decomposed into a series of primitive standard quantum logics, each implemented as narrowband microwave pulses applied to the quantum system. This standard approach suffers from cumulative errors of pulse concatenation and device decoherence error. The alternative approach is quantum optimal control, which computes a single microwave pulse that directly realizes the arbitrary quantum logic based on quantum optimal control theory. We present experimental demonstration of this approach for implementing highfidelity arbitrary quantum logic on a superconducting qudit. We describe our procedure for extracting the system Hamiltonian, calibrating the quantum and classical hardware chain, and evaluating the gate fidelity. 
Wednesday, March 17, 2021 5:36PM  5:48PM On Demand 
P33.00012: Reinforcement learning for errorrobust control on cloudbased superconducting hardware [Part I] Yuval Baum, Sean Howell, Maggie Liuzzi, Pranav Mundada, Mirko Amico, Michael Hush, Michael Biercuk The noisy nature of today's quantum hardware limits the ability to realize functioning quantum computers. Yet, a careful design of the systems' controls allows researchers to narrow the gap between current and desired hardware capabilities. In this work, we study a blackbox optimization technique based on reinforcement learning for the discovery of highperformance gatesets on a cloud quantum computer. We show that by employing RL, where intermediate information is used to optimize a long term goal, we are able to generate singlequbit gates which, when implemented on a real device, outperform existing modelbased optimized pulses. We demonstrate the performance of our learner on IBM quantum hardware accessed via Qiskit Pulse programming. The entire learning process occurs on the quantum device itself, aiming to suit the lowlevel gate implementation to the underlying details of the specific hardware. This allows gate optimization without any prior knowledge or assumptions on the noise model, hardware limitations or any other undesired effect that exists in real devices. Our experiments demonstrate gates up to 3 times faster than the IBM default, with performance of less than 3e4 EPG and stability up to 10 days, as compared to the singleday calibration window for IBM pulses. 
Wednesday, March 17, 2021 5:48PM  6:00PM On Demand 
P33.00013: Reinforcement learning for errorrobust control on cloudbased superconducting hardware [Part II] Mirko Amico, Yuval Baum, Sean Howell, Michael Hush, Maggie Liuzzi, Pranav Mundada, Michael Biercuk The noisy nature of today’s quantum hardware limits the ability to realize functioning and useful 
Follow Us 
Engage
Become an APS Member 
My APS
Renew Membership 
Information for 
About APSThe American Physical Society (APS) is a nonprofit 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 207403844
(301) 2093200
Editorial Office
1 Research Road, Ridge, NY 119612701
(631) 5914000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 200452001
(202) 6628700