Bulletin of the American Physical Society
APS March Meeting 2023
Volume 68, Number 3
Las Vegas, Nevada (March 5-10)
Virtual (March 20-22); Time Zone: Pacific Time
Session D58: Tensor Networks and Hybrid Quantum Algorithms |
Hide Abstracts |
Sponsoring Units: DQI Chair: Sarah Sheldon, IBM Quantum Room: Room 302 |
Monday, March 6, 2023 3:00PM - 3:12PM |
D58.00001: Simulating a Neutral Atom Open Quantum System with Tensor Network States James Z Allen, Matthew Otten, Stephen K Gray, Bryan K Clark There are many different computational frameworks that can simulate an open quantum system. In particular, tensor network states (TNS) excel for low-depth quantum circuits on a large number of sites. We model the density matrix of an open quantum system on a neutral atom array using different types of TNS, from matrix product operators (MPO) to matrix product density operators (MPDO). As the MPO by itself is not a faithful representation of a positive-definite matrix, we consider ways to improve its positivity without significantly increasing computation cost. To improve the performance of large tensor operations, we use a distributed-memory tensor contraction library, finding not only an improvement in wall time but also core-hour cost. Using this framework, we measure time evolution under the Transverse Field Ising Model (TFIM) to gauge the ability of a QAOA algorithm to optimize the TFIM through dissipation and dephasing noise. We also check the ability of an open quantum system to maintain accuracy under a randomly sampled circuit. |
Monday, March 6, 2023 3:12PM - 3:24PM |
D58.00002: Efficient tensor network simulation of quantum many-body physics on sparse graphs Subhayan Sahu, Brian Swingle We study tensor network states defined on an underlying graph which is sparsely connected. Generic sparse graphs are expander graphs with a high probability, and one can represent volume law entangled states efficiently with only polynomial resources. We find that message-passing inference algorithms such as belief propagation can lead to efficient computation of local expectation values for a class of tensor network states defined on sparse graphs. As applications, we study local properties of square root states, graph states, and also employ this method to variationally prepare ground states of gapped Hamiltonians defined on generic graphs. Using the variational method we study the phase diagram of the transverse field quantum Ising model defined on sparse expander graphs. |
Monday, March 6, 2023 3:24PM - 3:36PM |
D58.00003: Quantum Computing and HPC Integration Yuri Alexeev, Danylo Lykov, Minzhao Liu, William Berquist In this presentation, we will discuss how quantum computing (QC) and high-performance computing (HPC) complement each other. Various models for integration will be discussed. We will discuss how HPC can help QC and another way around how QC can help HPC. The emphasis will be made on the need to develop state-of-the-art quantum circuit simulators to run efficiently on modern supercomputers to run at scale large simulations. In particular, the development of tensor network simulators for GPU-based supercomputers. The main use cases for simulators are the design of new quantum algorithms and finding optimal parameters, verification of quantum supremacy and advantage, and co-design of QC+HPC architecture at both hardware and software system levels. We will present the current progress on the development of the various quantum simulators and future directions. |
Monday, March 6, 2023 3:36PM - 3:48PM |
D58.00004: Mapping multidimensional chemical dynamics problems to a family of hybrid quantum and classical computing environments Anurag Dwivedi, Miguel Angel Lopez-Ruiz, Debadrita Saha, Philip Richerme, Srinivasan S Iyengar Despite ubiquitous applications of chemical dynamics simulations, the rendered computational cost grows exponentially with the degrees of freedom of a chemical system, limiting a classical computer to only specific chemical systems. This "curse of dimensionality" is further compounded by the quantum treatment of nuclear degrees of freedom which requires accurate time evolution of nuclear wavefunctions on precisely computed potential energy surfaces. |
Monday, March 6, 2023 3:48PM - 4:00PM |
D58.00005: Preparation of Strongly Correlated States in Quantum Computer Byungmin Kang, Vito W Scarola, Kwon Park Solving interacting quantum systems is arguably one of the most important problems in condensed matter physics. Correlated electron states in cuprate materials and interacting electron gas in high magnetic field are prototypical examples where our theoretical understanding is rather limited. The main difficulty can often be captured by the so-called sign problem, which explains the inability of classical computers in describing many-body quantum states. Despite such difficulties, trial wave functions such as the resonating valence bond (RVB) states and the Laughlin state have proved to be successful in understanding the quantum many-body systems. Therefore, it is natural to expect that these wave functions would play important roles when studying correlated systems using quantum computers. In this talk, I will present how one can systematically prepare correlated many-body states in quantum circuits via recently developed amplitude amplification technique. Using the RVB state as our main example, I will explain the relevant parameters needed for building quantum circuits can all be efficiently computed in classical computers. This includes the estimation of the amplitude of our target RVB state in a Bardeen-Cooper-Schrieffer state, unlike in the general cases where the estimation of the target state in an initial state is challenging. I will conclude by discussing the generality of our approach in tackling strongly interacting systems using quantum computers. |
Monday, March 6, 2023 4:00PM - 4:12PM |
D58.00006: Large-Scale VQE Simulations with Tensor-Networks Abid A Khan, Norm M Tubman, Sohaib Alam, Bryan K Clark, Wayne Mullinax Approaches in quantum state preparation that include optimizing parameterized quantum circuits can be difficult due to noisy environments and barren plateaus, obscuring their utility as the number of qubits grow. Here, we show that purely classical resources can be used to optimize quantum circuits in an approximate but robust manner. Specifically, we approximate a parameterized circuit with a matrix product state (MPS) of a fixed bond dimension, and we find optimal parameters using classical solvers. We demonstrate this approach by parameterizing circuits representing ground states of the Hubbard model. By initializing parameterized quantum circuits with parameters obtained via classical optimization, we hope to avoid the many problems that occur with quantum algorithms. |
Monday, March 6, 2023 4:12PM - 4:24PM Author not Attending |
D58.00007: TensorQC: Towards Scalable Quantum Classical Hybrid Compute via Tensor Networks Wei Tang, Margaret Martonosi Quantum processing units (QPUs) have to satisfy highly demanding quantity and quality requirements on their qubits to produce accurate results for problems at useful scales. Furthermore, classical simulations of quantum circuits generally do not scale. Instead, quantum circuit cutting techniques cut and distribute a large quantum circuit into multiple smaller subcircuits feasible for less powerful QPUs. However, the classical post-processing incurred from the cutting introduces runtime and memory bottlenecks. This work presents Ten- sorQC, which addresses the bottlenecks via tensor network based post-processing that minimizes the classical overhead by orders of magnitudes over prior parallelization techniques. Our experiments reduce the quantum area requirement by at least 60% over the purely quantum platforms. We also demonstrated benchmarks up to 200 qubits on a single GPU, much beyond the reach of the purely classical platforms. |
Monday, March 6, 2023 4:24PM - 4:36PM |
D58.00008: Maximizing success and minimizing resources: An optimal design of hybrid algorithms for NISQ era devices Yulun Wang, Gavin Hartnett, Yuval Baum Hybrid algorithms can potentially deliver quantum advantage in problems from physics, chemistry and optimization. Practical implementations on current available quantum hardware are still challenged by optimization inefficiency, poor scalability, runtime overhead and inaccuracy. In this work, we redesign the QAOA and VQE algorithm systematically to efficiently improve the success rate and accuracy while using minimal resources. By automating and optimising the ansatz selection, cost function and parameters optimization we can reduce the complexity of the classical component of hybrid algorithms, leading to a much faster, stable and consistent convergence. We show the redesigned QAOA procedure exhibiting 6x improvement in success probability solving Max-Cut problem with at least 3 times less function calls compared to commonly used randomised initialization methods. We also demonstrate an improved accuracy and convergence speed from the redesigned VQE procedure. Finally, we demonstrate further algorithmic improvements achieved by applying our deterministic error-suppression workflow on NISQ hardware, which provides the hybrid algorithm a robust noise resistance and enables the scalability to larger devices. |
Monday, March 6, 2023 4:36PM - 4:48PM |
D58.00009: Entanglement regularization in the quantum dynamics of tensor network states Miguel Angel Lopez-Ruiz, Anurag Dwivedi, Srinivasan S Iyengar A well-known computational challenge in quantum many-body dynamics is the exponential scaling of the system degrees of freedom with its size. This so-called "course of dimensionality" renders the full description of an arbitrary quantum state virtually intractable due to the enormous dimension of the Hilbert space. In an effort to tackle this issue, tensor networks (TN) have emerged as a general data compression scheme which allow the truncation of the Hilbert space, by exploiting the intrinsic entanglement structure of quantum states. Moreover, with the emergent interest in quantum information, a demand for hybrid algorithms that are suitable for current and future NISQ devices is on the rise, and TN methods are considered to be compelling candidates. |
Monday, March 6, 2023 4:48PM - 5:00PM |
D58.00010: Time Evolution of Uniform Sequential Circuits Nikita Astrakhantsev, Sheng-Hsuan Lin, Adam Smith, Frank Pollmann Simulating time evolution of generic quantum many-body systems using classical numerical approaches has an exponentially growing cost either with evolution time or with the system size. In this work, we present a polynomially scaling hybrid quantum-classical algorithm for time evolving a one-dimensional uniform system in the thermodynamic limit. This algorithm uses a layered uniform sequential quantum circuit as a variational ansatz to represent infinite translation-invariant quantum states. We show numerically that this ansatz requires a number of parameters polynomial in the simulation time for a given accuracy. Furthermore, this favourable scaling of the ansatz is maintained during our variational evolution algorithm. All steps of the hybrid optimisation are designed with near-term digital quantum computers in mind. After benchmarking the evolution algorithm on a classical computer, we demonstrate the measurement of observables of this uniform state using a finite number of qubits on a cloud-based quantum processing unit. With more efficient tensor contraction schemes, this algorithm may also offer improvements as a classical numerical algorithm. |
Monday, March 6, 2023 5:00PM - 5:12PM |
D58.00011: Classical and Quantum Strategies to Boost Quantum Subspace Methods Daan Camps, Yizhi Shen, Katherine Klymko, Norm M Tubman, Roel Van Beeumen Quantum subspace methods are an exciting class of hybrid quantum algorithms for ground and excited state computations where approximate energies are extracted from an appropriate subspace of the full Hilbert space. The expansion states that form the basis for the subspace are prepared on the quantum computer and the projected problem is retrieved through measurement. The approximate energies are then obtained through classical diagonalization of the low-dimensional projected problem. In this talk, we present classical and quantum strategies that aim to improve the energy approximations and convergence behavior of quantum subspace algorithms by improving the conditioning of the basis of expansion states through both implicit and explicit methods. We show that our strategies lead to more accurate energies for comparable classical and quantum resources and illustrate the performance through numerical simulations for a variety of problems stemming from condensed matter physics and electronic structure theory. |
Monday, March 6, 2023 5:12PM - 5:24PM |
D58.00012: Many body Green's function using variational dynamics Niladri Gomes, Lindsay Bassman, David B Williams-Young, Wibe A de Jong We present a method to compute many-body real-time Green's function using adaptive variational quantum dynamics simulation. The real-time Green's function involves the time evolution of a quantum state with one additional electron w.r.t. the ground state wavefunction. Simulation of such a non-normal quantum state is achieved by expressing it as a linear combination of multiple branch states. The real-time evolution and Green's function are obtained by combining the dynamics of the individual branch states. In order to minimize the error of a convergent Fourier transform of the Green's function using finite time simulation, we use the Padé approximation of the real-time data. We apply our method to the Hubbard model at half-filling and find very good agreement with exact results. As a part of error mitigation, we develop a resolution-enhancing method that we successfully apply to noisy data. |
Monday, March 6, 2023 5:24PM - 5:36PM |
D58.00013: Quantum chemistry simulation of ground- and excited-state properties of the sulfonium cation on a superconducting quantum processor Mario Motta, Gavin O Jones, Julia E Rice, Tanvi P Gujarati, Rei Sakuma, Ieva Liepuoniute, Jeannette M Garcia, Yu-ya Ohnishi The computational description of correlated electronic structure, and particularly of excited states of many-electron systems, is an anticipated application for quantum devices. An important ramification is to determine the dominant molecular fragmentation pathways in photo-dissociation experiments of light-sensitive compounds, like sulfonium-based photo-acid generators used in photolithography. Here [1] we simulate the static and dynamical electronic structure of the sulfonium cation, taken as a minimal model of a triply bonded sulfur cation, on a superconducting quantum processor of the IBM Falcon architecture. |
Monday, March 6, 2023 5:36PM - 5:48PM |
D58.00014: Scaling and performance analysis of variational quantum algorithms for quantum simulation Mario Ponce Martinez, Inés de Vega, Martin Leib In the past decades, several algorithms for fault-tolerant quantum computers such as Berry’s algorithm for systems of ordinary differential equations [1], HHL for solving linear systems of equations [2] and Shor’s algorithm for factorization [3] have been proven to provide an advantage over classical methods. Their mathematical structure makes it possible to bound precision and asymptotic runtime scaling in terms of several key quantities such as problem size, allowed error and simulated time. On the other hand, variational quantum algorithms (VQA) [4] are one of the most promising classes of algorithms with the capability of bridging the gap between current NISQ- and future fault-tolerant (FT) quantum computers. However, because of their heuristic nature, VQA are considerably less amenable to performance evaluations and investigations of their asymptotic runtime scaling. In this talk we introduce an approximate numerical scaling analysis strategy for VQA for quantum simulation and show the results of applying it to a selection of cases. |
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