d-fine (Germany)

This paper introduces a novel approach to the Quantum Approximate Optimization Algorithm (QAOA), specifically tailored to the knapsack problem. We combine the recently proposed quantum tree generator as an efficient state preparation circuit for all feasible solutions to the knapsack problem with the framework of Grover-mixer QAOA to form the first representative of Amplitude Amplification-mixer QAOA (AAMQAOA). On hard benchmark sets with up to 20 knapsack items, we demonstrate our method's improved performance over the current state-of-the-art Copula-QAOA. However, for larger instance sizes, both approaches fail to deliver better outcomes than greedily packing items in descending value-to-weight ratio, at least for the considered circuit depths. For sufficiently high circuit depths, however, we can prove that AAM-QAOA will eventually be able to sample the optimal solution.

In this article we setup a dynamic device-to-device communication system where devices, given as a Poisson point process, move in an environment, given by a street system of random planar-tessellation type, via a random-waypoint model. Every device independently picks a target location on the street system using a general waypoint kernel, and travels to the target along the shortest path on the streets with an individual velocity. Then, any pair of devices becomes connected whenever they are on the same street in sufficiently close proximity, for a sufficiently long time. In our main results we isolate regimes for the almost-sure absence of percolation if, for example, the device intensity is too small, or the connectivity time is too large. On the other hand, we exhibit parameter regimes of sufficiently large intensities of devices, under favorable choices of the other parameters, such that percolation is possible with positive probability. Most interestingly, we also show an in-and-out of percolation as the velocity increases. The rigorous analysis of the system mainly rests on comparison arguments with simplified models via spatial coarse graining and thinning approaches. Here we also make contact to geostatistical percolation models with infinite-range dependencies.

The study of selected logs from the Central-Eastern Iran Microplate (CEIM; area of Arusan) and Sanandaj–Sirjan Zone (SSZ; areas of Esfahan east and Tiran) constrains timing and evolution of the Cretaceous successions, resting unconformably on older sediments or non-conformably on the metamorphic basement. The Cretaceous marine transgression occurred almost synchronously on a levelled topography, documenting a previous deep erosion. Above a basal siliciclastic unit, rudist- and orbitolinid-rich lower Aptian limestone (“Orbitolina Limestone”, Shak Kuh Formation, Taft Formation) were deposited. Later, the stratigraphic evolution of the three studied areas differentiates, documenting changes in the depositional environments, reflecting different subsidence rates. In the Tiran area, marls and cherty limestones covering shallow-water limestones record a deepening trend. A differentiation of depositional environment is observed at Arusan, where conglomerates to the west pass to basinal sandstone and marls to the east. Here, a Coniacian-Campanian reprise of the carbonate production (with skeletal limestones rich in rudists, ostreids, and abundant orbitoidids) is covered by cherty marls. In the Esfahan north area, glauconitic hardground marls covered by chert record important deepening during the early Cenomanian, followed by the reprise of shallow-water carbonate production (middle Cenomanian). The synchronous deposition of Aptian carbonate successions on large parts of the CEIM, SSZ and North-Iranian domains suggests a rapid regional transgression controlled by a sea-level rise. Stratigraphic differentiation occurring in late Aptian–Cenomanian times, recording the onset of tectonic events differently affecting the aforementioned domains, likely related to the opening of oceanic seaways in the CEIM and the SSZ.

Generative adversarial networks (GANs) are a machine learning framework comprising a generative model for sampling from a target distribution and a discriminative model for evaluating the proximity of a sample to the target distribution. GANs exhibit strong performance in imaging or anomaly detection. However, they suffer from training instabilities, and sampling efficiency may be limited by the classical sampling procedure. We introduce variational quantum–classical Wasserstein GANs (WGANs) to address these issues and embed this model in a classical machine learning framework for anomaly detection. Classical WGANs improve training stability by using a cost function better suited for gradient descent. Our model replaces the generator of WGANs with a hybrid quantum–classical neural net and leaves the classical discriminative model unchanged. This way, high-dimensional classical data only enters the classical model and need not be prepared in a quantum circuit. We demonstrate the effectiveness of this method on a credit card fraud dataset. For this dataset our method shows performance on par with classical methods in terms of the F 1 score. We analyze the influence of the circuit ansatz, layer width and depth, neural net architecture parameter initialization strategy, and sampling noise on convergence and performance.

Given the increasing interest in and the growing number of publicly available methods to estimate investor sentiment from social media platforms, researchers and practitioners alike are facing one crucial question – which is best to gauge investor sentiment? We compare the performance of daily investor sentiment measures estimated from Twitter and StockTwits short messages by publicly available dictionary and machine learning based methods for a large sample of stocks. To determine their relevance for financial applications, these investor sentiment measures are compared by their effects on the cross-section of stocks (i) within a Fama and MacBeth (J Polit Econ 81:607–636, 1973) regression framework applied to a measure of retail investors’ order imbalances and (ii) by their ability to forecast abnormal returns in a model-free portfolio sorting exercise. Interestingly, we find that investor sentiment measures based on finance-specific dictionaries do not only have a greater impact on retail investors’ order imbalances than measures based on machine learning approaches, but also perform very well compared to the latter in our asset pricing application.

A new Bateman-Hillion solution to the Dirac equation for a relativistic Gaussian electron beam taking explicit account of the four-position of the beam waist is presented. This solution has a pure Gaussian form in the paraxial limit but beyond it contains higher order Laguerre-Gaussian components attributable to the tighter focusing. One implication of the mixed mode nature of strongly diffracting beams is that the expectation values for spin and orbital angular momenta are fractional and are interrelated to each other by intrinsic spin-orbit coupling. Our results for these properties align with earlier work on Bessel beams [Bliokh etal., Phys. Rev. Lett. 107, 174802 (2011)PRLTAO0031-900710.1103/PhysRevLett.107.174802] and show that fractional angular momenta can be expressed by means of a Berry phase. The most significant difference arises, though, due to the fact that Laguerre-Gaussian beams naturally contain Gouy phase, while Bessel beams do not. We show that Gouy phase is also related to Berry phase and that Gouy phase fronts that are flat in the paraxial limit become curved beyond it.

BackgroundMistletoe extracts are used as an adjunct therapy for cancer patients, but there is dissent as to whether this therapy has a positive impact on quality of life (QoL).MethodsWe conducted a systematic review searching in several databases (Medline, Embase, CENTRAL, CINAHL, PsycInfo, Science Citation Index, clinicaltrials.gov, opengrey.org) by combining terms that cover the fields of “neoplasm”, “quality of life” and “mistletoe”. We included prospective controlled trials that compared mistletoe extracts with a control in cancer patients and reported QoL or related dimensions. The quality of the studies was assessed with the Cochrane Risk of Bias tool version 2. We conducted a quantitative meta-analysis.ResultsWe included 26 publications with 30 data sets. The studies were heterogeneous. The pooled standardized mean difference (random effects model) for global QoL after treatment with mistletoe extracts vs. control was d = 0.61 (95% CI 0.41–0.81, p < 0,00001). The effect was stronger for younger patients, with longer treatment, in studies with lower risk of bias, in randomized and blinded studies. Sensitivity analyses support the validity of the finding. 50% of the QoL subdomains (e.g. pain, nausea) show a significant improvement after mistletoe treatment. Most studies have a high risk of bias or at least raise some concern.ConclusionMistletoe extracts produce a significant, medium-sized effect on QoL in cancer. Risk of bias in the analyzed studies is likely due to the specific type of treatment, which is difficult to blind; yet this risk is unlikely to affect the outcome.PROSPERO registrationCRD42019137704.