Improved machine learning algorithm for predicting ground state properties

Finding the ground state of a quantum many-body system is a fundamental problem in quantum physics. In this work, we give a classical machine learning (ML) algorithm for predicting ground state properties with an inductive bias encoding geometric locality. The proposed ML model can efficiently predict ground state properties of an n-qubit gapped local Hamiltonian after learning from only [Formula: see text] data about other Hamiltonians in the same quantum phase of matter. This improves substantially upon previous results that require [Formula: see text] data for a large constant c. Furthermore, the training and prediction time of the proposed ML model scale as [Formula: see text] in the number of qubits n. Numerical experiments on physical systems with up to 45 qubits confirm the favorable scaling in predicting ground state properties using a small training dataset.

Provably efficient machine learning for quantum many-body problems

Classical machine learning (ML) provides a potentially powerful approach to solving challenging quantum many-body problems in physics and chemistry. However, the advantages of ML over traditional methods have not been firmly established. In this work, we prove that classical ML algorithms can efficiently predict ground-state properties of gapped Hamiltonians after learning from other Hamiltonians in the same quantum phase of matter. By contrast, under a widely accepted conjecture, classical algorithms that do not learn from data cannot achieve the same guarantee. We also prove that classical ML algorithms can efficiently classify a wide range of quantum phases. Extensive numerical experiments corroborate our theoretical results in a variety of scenarios, including Rydberg atom systems, two-dimensional random Heisenberg models, symmetry-protected topological phases, and topologically ordered phases.

Quantum advantage in learning from experiments

Quantum technology promises to revolutionize how we learn about the physical world. An experiment that processes quantum data with a quantum computer could have substantial advantages over conventional experiments in which quantum states are measured and outcomes are processed with a classical computer. We proved that quantum machines could learn from exponentially fewer experiments than the number required by conventional experiments. This exponential advantage is shown for predicting properties of physical systems, performing quantum principal component analysis, and learning about physical dynamics. Furthermore, the quantum resources needed for achieving an exponential advantage are quite modest in some cases. Conducting experiments with 40 superconducting qubits and 1300 quantum gates, we demonstrated that a substantial quantum advantage is possible with today's quantum processors.

Efficient Estimation of Pauli Observables by Derandomization

We consider the problem of jointly estimating expectation values of many Pauli observables, a crucial subroutine in variational quantum algorithms. Starting with randomized measurements, we propose an efficient derandomization procedure that iteratively replaces random single-qubit measurements by fixed Pauli measurements; the resulting deterministic measurement procedure is guaranteed to perform at least as well as the randomized one. In particular, for estimating any L low-weight Pauli observables, a deterministic measurement on only of order log(L) copies of a quantum state suffices. In some cases, for example, when some of the Pauli observables have high weight, the derandomized procedure is substantially better than the randomized one. Specifically, numerical experiments highlight the advantages of our derandomized protocol over various previous methods for estimating the ground-state energies of small molecules.

Information-Theoretic Bounds on Quantum Advantage in Machine Learning

We study the performance of classical and quantum machine learning (ML) models in predicting outcomes of physical experiments. The experiments depend on an input parameter x and involve execution of a (possibly unknown) quantum process E. Our figure of merit is the number of runs of E required to achieve a desired prediction performance. We consider classical ML models that perform a measurement and record the classical outcome after each run of E, and quantum ML models that can access E coherently to acquire quantum data; the classical or quantum data are then used to predict the outcomes of future experiments. We prove that for any input distribution D(x), a classical ML model can provide accurate predictions on average by accessing E a number of times comparable to the optimal quantum ML model. In contrast, for achieving an accurate prediction on all inputs, we prove that the exponential quantum advantage is possible. For example, to predict the expectations of all Pauli observables in an n-qubit system ρ, classical ML models require 2^{Ω(n)} copies of ρ, but we present a quantum ML model using only O(n) copies. Our results clarify where the quantum advantage is possible and highlight the potential for classical ML models to address challenging quantum problems in physics and chemistry.

TriB-RT: Simultaneous optimization of photon, electron and proton beams

PURPOSE

To develop a novel treatment planning process (TPP) with simultaneous optimization of modulated photon, electron and proton beams for improved treatment plan quality in radiotherapy.

METHODS

A framework for fluence map optimization of Monte Carlo (MC) calculated beamlet dose distributions is developed to generate treatment plans consisting of photon, electron and spot scanning proton fields. Initially, in-house intensity modulated proton therapy (IMPT) plans are compared to proton plans created by a commercial treatment planning system (TPS). A triple beam radiotherapy (TriB-RT) plan is generated for an exemplary academic case and the dose contributions of the three particle types are investigated. To investigate the dosimetric potential, a TriB-RT plan is compared to an in-house IMPT plan for two clinically motivated cases. Benefits of TriB-RT for a fixed proton beam line with a single proton field are investigated.

RESULTS

In-house optimized IMPT are of at least equal or better quality than TPS-generated proton plans, and MC-based optimization shows dosimetric advantages for inhomogeneous situations. Concerning TriB-RT, for the academic case, the resulting plan shows substantial contribution of all particle types. For the clinically motivated case, improved sparing of organs at risk close to the target volume is achieved compared to IMPT (e.g. myelon and brainstem [Formula: see text] -37%) at cost of an increased low dose bath (healthy tissue V _10% +22%). In the scenario of a fixed proton beam line, TriB-RT plans are able to compensate the loss in degrees of freedom to substantially improve plan quality compared to a single field proton plan.

CONCLUSION

A novel TPP which simultaneously optimizes photon, electron and proton beams was successfully developed. TriB-RT shows the potential for improved treatment plan quality and is especially promising for cost-effective single-room proton solutions with a fixed beamline in combination with a conventional linac delivering photon and electron fields.

Mixed-State Entanglement from Local Randomized Measurements

We propose a method for detecting bipartite entanglement in a many-body mixed state based on estimating moments of the partially transposed density matrix. The estimates are obtained by performing local random measurements on the state, followed by postprocessing using the classical shadows framework. Our method can be applied to any quantum system with single-qubit control. We provide a detailed analysis of the required number of experimental runs, and demonstrate the protocol using existing experimental data [Brydges et al., Science 364, 260 (2019)SCIEAS0036-807510.1126/science.aau4963].

Development of an extended Macro Monte Carlo method for efficient and accurate dose calculation in magnetic fields

MOTIVATION

Progress in the field of magnetic resonance (MR)-guided radiotherapy has triggered the need for fast and accurate dose calculation in presence of magnetic fields. The aim of this work is to satisfy this need by extending the macro Monte Carlo (MMC) method to enable dose calculation for photon, electron, and proton beams in a magnetic field.

METHODS

The MMC method is based on the transport of particles in macroscopic steps through an absorber by sampling the relevant physical quantities from a precalculated database containing probability distribution functions. To enable MMC particle transport in a magnetic field, a transformation accounting for the Lorentz force is applied for each macro step by rotating the sampled position and direction around the magnetic field vector. The transformed position and direction distributions on local geometries are validated against full MC for electron and proton pencil beams. To enable photon dose calculation, an in-house MC algorithm is used for photon transport and interaction. Emerging secondary charged particles are passed to MMC for transport and energy deposition. The extended MMC dose calculation accuracy and efficiency is assessed by comparison with EGSnrc (photon and electron beams) and Geant4 (proton beam) calculated dose distributions of different energies and homogeneous magnetic fields for broad beams impinging on water phantoms with bone and lung inhomogeneities.

RESULTS

The geometric transformation on the local geometries is able to reproduce the results of full MC for all investigated settings (difference in mean value and standard deviation <1%). Macro Monte Carlo calculated dose distributions in a homogeneous magnetic field are in agreement with EGSnrc and Geant4, respectively, with gamma passing rates >99.6% (global 2%, 2 mm and 10% threshold criteria) for all situations. MMC achieves a substantial efficiency gain of up to a factor of 21 (photon beam), 66 (electron beam), and 356 (proton beam) compared to EGSnrc or Geant4.

CONCLUSION

Efficient and accurate dose calculation in magnetic fields was successfully enabled by utilizing the developed extended MMC transport method for photon, electron, and proton beams.

Recovering Quantum Gates from Few Average Gate Fidelities

Characterizing quantum processes is a key task in the development of quantum technologies, especially at the noisy intermediate scale of today's devices. One method for characterizing processes is randomized benchmarking, which is robust against state preparation and measurement errors and can be used to benchmark Clifford gates. Compressed sensing techniques achieve full tomography of quantum channels essentially at optimal resource efficiency. In this Letter, we show that the favorable features of both approaches can be combined. For characterizing multiqubit unitary gates, we provide a rigorously guaranteed and practical reconstruction method that works with an essentially optimal number of average gate fidelities measured with respect to random Clifford unitaries. Moreover, for general unital quantum channels, we provide an explicit expansion into a unitary 2-design, allowing for a practical and guaranteed reconstruction also in that case. As a side result, we obtain a new statistical interpretation of the unitarity-a figure of merit characterizing the coherence of a process.

Part 2: Dynamic mixed beam radiotherapy (DYMBER): Photon dynamic trajectories combined with modulated electron beams

PURPOSE

The purpose of this study was to develop a treatment technique for dynamic mixed beam radiotherapy (DYMBER) utilizing increased degrees of freedom (DoF) of a conventional treatment unit including different particle types (photons and electrons), intensity and energy modulation and dynamic gantry, table, and collimator rotations.

METHODS

A treatment planning process has been developed to create DYMBER plans combining photon dynamic trajectories (DTs) and step and shoot electron apertures collimated with the photon multileaf collimator (pMLC). A gantry-table path is determined for the photon DTs with minimized overlap of the organs at risk (OARs) with the target. In addition, an associated dynamic collimator rotation is established with minimized area between the pMLC leaves and the target contour. pMLC sequences of photon DTs and electron pMLC apertures are then simultaneously optimized using direct aperture optimization (DAO). Subsequently, the final dose distribution of the electron pMLC apertures is calculated using the Swiss Monte Carlo Plan (SMCP). The pMLC sequences of the photon DTs are then re-optimized with a finer control point resolution and with the final electron dose distribution taken into account. Afterwards, the final photon dose distribution is calculated also using the SMCP and summed together with the one of the electrons. This process is applied for a brain and two head and neck cases. The resulting DYMBER dose distributions are compared to those of dynamic trajectory radiotherapy (DTRT) plans consisting only of photon DTs and clinically applied VMAT plans. Furthermore, the deliverability of the DYMBER plans is verified in terms of dosimetric accuracy, delivery time and collision avoidance. For this purpose, The DYMBER plans are delivered to Gafchromic EBT3 films placed in an anthropomorphic head phantom on a Varian TrueBeam linear accelerator.

RESULTS

For each case, the dose homogeneity in the target is similar or better for DYMBER compared to DTRT and VMAT. Averaged over all three cases, the mean dose to the parallel OARs is 16% and 28% lower, D2% to the serial OARs is 17% and 37% lower and V10% to normal tissue is 12% and 4% lower for the DYMBER plans compared to the DTRT and VMAT plans, respectively. The DYMBER plans are delivered without collision and with a 4-5 min longer delivery time than the VMAT plans. The absolute dose measurements are compared to calculation by gamma analysis using 2% (global)/2 mm criteria with passing rates of at least 99%.

CONCLUSIONS

A treatment technique for DYMBER has been successfully developed and verified for its deliverability. The dosimetric superiority of DYMBER over DTRT and VMAT indicates utilizing increased DoF to be the key to improve brain and head and neck radiation treatments in future.

Comparing Experiments to the Fault-Tolerance Threshold

Achieving error rates that meet or exceed the fault-tolerance threshold is a central goal for quantum computing experiments, and measuring these error rates using randomized benchmarking is now routine. However, direct comparison between measured error rates and thresholds is complicated by the fact that benchmarking estimates average error rates while thresholds reflect worst-case behavior when a gate is used as part of a large computation. These two measures of error can differ by orders of magnitude in the regime of interest. Here we facilitate comparison between the experimentally accessible average error rates and the worst-case quantities that arise in current threshold theorems by deriving relations between the two for a variety of physical noise sources. Our results indicate that it is coherent errors that lead to an enormous mismatch between average and worst case, and we quantify how well these errors must be controlled to ensure fair comparison between average error probabilities and fault-tolerance thresholds.

Unifying framework for relaxations of the causal assumptions in Bell's theorem

Bell's theorem shows that quantum mechanical correlations can violate the constraints that the causal structure of certain experiments impose on any classical explanation. It is thus natural to ask to which degree the causal assumptions-e.g., locality or measurement independence-have to be relaxed in order to allow for a classical description of such experiments. Here we develop a conceptual and computational framework for treating this problem. We employ the language of Bayesian networks to systematically construct alternative causal structures and bound the degree of relaxation using quantitative measures that originate from the mathematical theory of causality. The main technical insight is that the resulting problems can often be expressed as computationally tractable linear programs. We demonstrate the versatility of the framework by applying it to a variety of scenarios, ranging from relaxations of the measurement independence, locality, and bilocality assumptions, to a novel causal interpretation of Clauser-Horne-Shimony-Holt inequality violations.