A5052211779- Quantum Computing Algorithms and Architecture
- Quantum Information and Cryptography
- Quantum and electron transport phenomena
- Quantum Mechanics and Applications
- Parallel Computing and Optimization Techniques
- Quantum-Dot Cellular Automata
- Philosophy and History of Science
- Cryptography and Data Security
- Advanced Thermodynamics and Statistical Mechanics
- Neural Networks and Applications
- Atomic and Subatomic Physics Research
- advanced mathematical theories
- Quantum many-body systems
- Advancements in Semiconductor Devices and Circuit Design
- Distributed and Parallel Computing Systems
- Neural Networks and Reservoir Computing
- Advanced Optimization Algorithms Research
- Cold Atom Physics and Bose-Einstein Condensates
- Cloud Computing and Resource Management
- Quantum, superfluid, helium dynamics
- Machine Learning and Algorithms
- Sensor Technology and Measurement Systems
- Industrial Technology and Control Systems
- Advanced Electrical Measurement Techniques
- Nuclear Physics and Applications
Forschungszentrum Jülich
2017-2025
RWTH Aachen University
2018-2020
We benchmark the quantum processing units of largest annealers to date, 5000+ qubit annealer Advantage and its 2000+ predecessor D-Wave 2000Q, using tail assignment exact cover problems from aircraft scheduling scenarios. The set contains small, intermediate, large with both sparsely connected almost fully instances. find that outperforms 2000Q for all problems, a notable increase in success rate problem size. In particular, is also able solve 120 logical qubits cannot anymore. Furthermore,...
Abstract Approaches to developing large-scale superconducting quantum processors must cope with the numerous microscopic degrees of freedom that are ubiquitous in solid-state devices. State-of-the-art qubits employ aluminium oxide (AlO x ) tunnel Josephson junctions as sources nonlinearity necessary perform operations. Analyses these typically assume an idealized, purely sinusoidal current–phase relation. However, this relation is expected hold only limit vanishingly low-transparency...
Kernel-based support vector machines (SVMs) are supervised machine learning algorithms for classification and regression problems. We introduce a method to train SVMs on D-Wave 2000Q quantum annealer study its performance in comparison trained conventional computers. The is applied both synthetic data real obtained from biology experiments. find that the produces an ensemble of different solutions often generalizes better unseen than single global minimum SVM computer, especially cases where...
Abstract The performance of the quantum approximate optimization algorithm is evaluated by using three different measures: probability finding ground state, energy expectation value, and a ratio closely related to approximation ratio. set problem instances studied consists weighted MaxCut problems 2-satisfiability problems. Ising model representations latter possess unique states highly degenerate first excited states. executed on computer simulators IBM Q Experience. Additionally, data...
A revised version of the massively parallel simulator a universal quantum computer, described in this journal eleven years ago, is used to benchmark various gate-based algorithms on some most powerful supercomputers that exist today. Adaptive encoding wave function reduces memory requirement by factor eight, making it possible simulate computers with up 48 qubits Sunway TaihuLight and K computer. The exhibits close-to-ideal weak-scaling behavior TaihuLight, an IBM Blue Gene/Q, Intel Xeon...
With the advent of public access to small gate-based quantum processors, it becomes necessary develop a benchmarking methodology such that independent researchers can validate operation these processors. We explore usefulness number simple circuits as benchmarks for computing devices and show performing identity operations are very simple, scalable sensitive gate errors therefore well suited this task. illustrate procedure by presenting benchmark results IBM Quantum Experience, cloud-based...
Abstract Solving combinatorial optimization problems of the kind that can be codified by quadratic unconstrained binary (QUBO) is a promising application quantum computation. Some this class suitable for practical applications such as traveling salesman problem (TSP), bin packing (BPP), or knapsack (KP) have inequality constraints require particular cost function encoding. The common approach use slack variables to represent in function. However, considerably increases number qubits and...
In the model of gate-based quantum computation, qubits are controlled by a sequence gates. superconducting qubit systems, these gates can be implemented voltage pulses. The success implementing particular gate expressed various metrics such as average fidelity, diamond distance, and unitarity. We analyze pulses for system two transmon coupled resonator, inspired architecture IBM Quantum Experience. obtained numerical solution time-dependent Schr\"odinger equation system. find that reflect...
We review and extend, in a self-contained way, the mathematical foundations of numerical simulation methods that are based on use random states. The power versatility this technology is illustrated by calculations physically relevant properties such as density states large single particle systems, specific heat, current–current correlations, density–density electron spin resonance spectra many-body systems. explore new field applications state showing it can be used to analyze simulations...
Support Vector Machine (SVM) is a popular supervised Learning (ML) method that widely used for classification and regression problems. Recently, to train SVMs on D-Wave 2000Q Quantum Annealer (QA) was proposed binary of some biological data. First, ensembles weak quantum are generated by training each classifier disjoint subset can be fit into the QA. Then, computed solutions fused making predictions unseen In this work, Remote Sensing (RS) multispectral images with trained QA discussed....
Reading out the state of superconducting artificial atoms typically relies on dispersive coupling to a readout resonator. For given system noise temperature, increasing circulating photon number $\bar{n}$ in resonator enables shorter measurement time and is therefore expected reduce errors caused by spontaneous atom transitions. However, generally observed also increase these transition rates. Here we present fluxonium which measure an overall flat dependence rates between its first two...
We study large-scale applications using a GPU-accelerated version of the massively parallel J\"ulich universal quantum computer simulator (JUQCS--G). First, we benchmark JUWELS Booster, GPU cluster with 3744 NVIDIA A100 Tensor Core GPUs. Then, use JUQCS--G to relation between annealing (QA) and approximate optimization algorithm (QAOA). find that very coarsely discretized QA, termed (AQA), performs surprisingly well in comparison QAOA. It can either be used initialize QAOA, or avoid costly...
We utilize the theory of local amplitude transfer (LAT) to gain insights into quantum walks (QWs) and annealing (QA) beyond adiabatic theorem. By representing eigenspace problem Hamiltonian as a hypercube graph, we demonstrate that probability traverses search space through series Rabi oscillations. argue movement can be systematically guided towards ground state using time-dependent hopping rate based solely on problem's energy spectrum. Building upon these insights, extend concept...
The analysis of empirical data through model-free inequalities leads to the conclusion that violations Bell-type by cannot have any significance unless one believes universe operates according rules a mathematical model.
Shor’s factoring algorithm is one of the most anticipated applications quantum computing. However, limited capabilities today’s computers only permit a study for very small numbers. Here, we show how large GPU-based supercomputers can be used to assess performance numbers that are out reach current and near-term hardware. First, original algorithm. While theoretical bounds suggest success probabilities 3–4%, find average above 50%, due high frequency “lucky” cases, defined as successful...
We benchmark the 5000+ qubit system Advantage coupled with Hybrid Solver Service 2 released by D-Wave Systems Inc. in September 2020 using a new class of optimization problems called garden known companion planting. These are scalable to an arbitrarily large number variables and intuitively find application real-world scenarios. derive their QUBO formulation illustrate relation quadratic assignment problem. demonstrate that hybrid solver can solve larger less time than predecessors. However,...
We take the point of view that building a one-way bridge from experimental data to mathematical models instead other way around avoids running into controversies resulting attaching meaning symbols used in latter. In particular, we show adopting this offers new perspectives for constructing and interpreting results Einstein-Podolsky-Rosen-Bohm experiments. first prove Bell-type inequalities constraining values four correlations obtained by performing experiments under different conditions....
Solving combinatorial optimization problems (COPs) is a promising application of quantum computation, with the Quantum Approximate Optimization Algorithm (QAOA) being one most studied algorithms for solving them. However, multiple factors make parameter search QAOA hard problem. In this work, we study transfer learning (TL), methodology to reuse pre-trained parameters problem instance into different COP instances. To end, select small cases traveling salesman (TSP), bin packing (BPP),...
We extensively test a recent protocol to demonstrate quantum fault tolerance on three systems: (1) real-time simulation of five spin qubits coupled an environment with two-level defects, (2) transmon computers, and (3) the 16-qubit processor IBM Q Experience. In simulations, dynamics full system is obtained by numerically solving time-dependent Schr\"odinger equation. find that fault-tolerant scheme provides systematic way improve results when errors are dominated inherent control...
Solving combinatorial optimization problems of the kind that can be codified by quadratic unconstrained binary (QUBO) is a promising application quantum computation. Some this class suitable for practical applications such as traveling salesman problem (TSP), bin packing (BPP), or knapsack (KP) have inequality constraints require particular cost function encoding. The common approach use slack variables to represent in function. However, considerably increases number qubits and operations...
We use discrete-event simulation to construct a subquantum model that can reproduce the quantum-theoretical prediction for statistics of data produced by Einstein-Podolsky-Rosen-Bohm experiment and an extension thereof. This satisfies Einstein's criterion locality generates in event-by-event cause-and-effect manner. show quantum theory describe certain range parameters only.
Benchmarks are essential in the design of modern HPC installations, as they define key aspects system components. Beyond synthetic workloads, it is crucial to include real applications that represent user requirements into benchmark suites, guarantee high usability and widespread adoption a new system. Given significant investments leadership-class supercomputers exascale era, this even more important necessitates alignment with vision Open Science reproducibility. In work, we present...
We introduce the notion of "separation conditions" meaning that a description statistical data obtained from experiments, performed under set different conditions, allows for decomposition such each partial depends on mutually exclusive subsets these conditions. Descriptions allow separation conditions are shown to entail basic mathematical framework quantum theory. The Stern-Gerlach and Einstein-Podolsky-Rosen-Bohm experiment with three, respectively nine possible outcomes used illustrate...