QInfer: Statistical inference software for quantum applications

Granade, Christopher; Ferrie, Christopher; Hincks, Ian; Casagrande, Steven; Alexander, Thomas; Gross, Jonathan A.; Kononenko, Michal; Sanders, Yuval R.

doi:10.22331/q-2017-04-25-5

Cited by 47 publications

(60 citation statements)

References 63 publications

(94 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The numerical methods used in this paper are implemented using the QInfer, QuaEC, QuTiP 2 and SciPy libraries for Python [43,[50][51][52]. We thank Holger Haas for simulating the gate sets used above, and for discussions.…”

Section: Acknowledgmentsmentioning

confidence: 99%

Accelerated randomized benchmarking

2015

Self Cite

View full text Add to dashboard Cite

Quantum information processing offers promising advances for a wide range of fields and applications, provided that we can efficiently assess the performance of the control applied in candidate systems. That is, we must be able to determine whether we have implemented a desired gate, and refine accordingly. Randomized benchmarking reduces the difficulty of this task by exploiting symmetries in quantum operations. Here, we bound the resources required for benchmarking and show that, with prior information, we can achieve several orders of magnitude better accuracy than in traditional approaches to benchmarking. Moreover, by building on state-of-the-art classical algorithms, we reach these accuracies with near-optimal resources. Our approach requires an order of magnitude less data to achieve the same accuracies and to provide online estimates of the errors in the reported fidelities. We also show that our approach is useful for physical devices by comparing to simulations.Quantum information processing devices offer great promise in a variety of different fields, including chemistry and material science, data analysis and machine learning [1][2][3][4], as well as cryptography [5]. Over the past few years, proposals have been advanced for quantum information processing past the classical scale, based on nodebased architectures [6,7]. In addition, rapid progress has been made towards experimental implementations that might allow for developing such devices [8,9]. An impediment in this effort, however, is presented by the difficulty of calibrating and diagnosing quantum devices.In particular, in the development of quantum information processing, an important experimental challenge is to efficiently characterize the quality with which we can control a quantum system. By characterizing the quality of a quantum gate that is implemented by a control pulse, we can then reason about the utility of that gate for quantum information processing tasks. For instance, we can estimate the feasibility of and the resources required to implement error correction using that control by comparing to proven and numerically estimated fault-tolerance thresholds [10,11]. Alternately, we can adjust our control sequences to account for differences between our control model and the actual system.In cases where only the quality of a quantum gate or set of gates is required, randomized benchmarking has proven to be a useful means of extracting this information with relatively little experimental effort [12]. This has been demonstrated in a variety of experimental settings [9,[13][14][15][16][17][18][19]. Randomized benchmarking has also been used to improve gate fidelities by characterizing cross-talk [20] or distortions [21]. Extracting fidelity information can often be useful in diagnosing performance and problems with a device in lieu of full characterization [22]. Moreover, randomized benchmarking has also been used to extract information about the completely positive and unital parts of linear maps [23].Here, using near-optimal data proces...

show abstract

Section: Acknowledgmentsmentioning

confidence: 99%

Accelerated randomized benchmarking

2015

Self Cite

View full text Add to dashboard Cite

show abstract

“…Taking a convex hull over such a region then provides a region which naturally includes the convexity of state space, and a minimum-volume enclosing ellipsoid over a credible region yields a compact description [25]. Both of these credible region estimators are included in the open-source package that we rely on for numerical implementations, QInfer [39]. Thus, we inherit a variety of practical data-driven credible region estimators.…”

Section: Bayesian Tomographymentioning

confidence: 99%

“…To address the issue of adoption we implement SMC-based tomography using an open-source library for Python [39] that integrates with the widely used QuTiP library for quantum information [40], and can be used with modern instrument control software [41]. The techniques that we introduce in this paper can readily be applied in experimental practice-we provide a tutorial on software implementations in appendix F.…”

Section: Introductionmentioning

confidence: 99%

Practical Bayesian tomography

2016

Self Cite

View full text Add to dashboard Cite

In recent years, Bayesian methods have been proposed as a solution to a wide range of issues in quantum state and process tomography. State-of-the-art Bayesian tomography solutions suffer from three problems: numerical intractability, a lack of informative prior distributions, and an inability to track time-dependent processes. Here, we address all three problems. First, we use modern statistical methods, as pioneered by Huszár and Houlsby (2012 Phys. Rev. A 85 052120) and by Ferrie (2014 New J. Phys.16 093035), to make Bayesian tomography numerically tractable. Our approach allows for practical computation of Bayesian point and region estimators for quantum states and channels. Second, we propose the first priors on quantum states and channels that allow for including useful experimental insight. Finally, we develop a method that allows tracking of time-dependent states and estimates the drift and diffusion processes affecting a state. We provide source code and animated visual examples for our methods.

show abstract

“…The numerical examples shown in sections 4, 5 and appendix A were obtained using an implementation in Python 2.7 (Anaconda distribution), with the NumPy [32], SciPy [33], and QInfer [34]…”

Section: { } (Set Of Integers)mentioning

confidence: 99%

Structured filtering

Granade

Wiebe

2017

New J. Phys.

Self Cite

View full text Add to dashboard Cite

A major challenge facing existing sequential Monte Carlo methods for parameter estimation in physics stems from the inability of existing approaches to robustly deal with experiments that have different mechanisms that yield the results with equivalent probability. We address this problem here by proposing a form of particle filtering that clusters the particles that comprise the sequential Monte Carlo approximation to the posterior before applying a resampler. Through a new graphical approach to thinking about such models, we are able to devise an artificial-intelligence based strategy that automatically learns the shape and number of the clusters in the support of the posterior. We demonstrate the power of our approach by applying it to randomized gap estimation and a form of low circuit-depth phase estimation where existing methods from the physics literature either exhibit much worse performance or even fail completely.

show abstract

QInfer: Statistical inference software for quantum applications

Cited by 47 publications

References 63 publications

Accelerated randomized benchmarking

Accelerated randomized benchmarking

Practical Bayesian tomography

Structured filtering

Contact Info

Product

Resources

About