Search

Observations of polarization position angle ($\theta$) standards made from 2014 to 2023 with the High Precision Polarimetric Instrument (HIPPI) and other HIPPI-class polarimeters in both hemispheres are used to investigate their variability. Multi-band data were first used to thoroughly recalibrate the instrument performance by bench-marking against carefully selected literature data. A novel Co-ordinate Difference Matrix (CDM) approach - which combines pairs of points - was then used to amalgamate monochromatic ($g^\prime$ band) observations from many observing runs and re-determine $\theta$ for 17 standard stars. The CDM algorithm was then integrated into a fitting routine and used to establish the impact of stellar variability on the measured position angle scatter. The approach yields variability detections for stars on long time scales that appear stable over short runs. The best position angle standards are $\ell$ Car, $o$ Sco, HD 154445, HD 161056 and $\iota^1$ Sco which are stable to $\leq$ 0.123$^\circ$. Position angle variability of 0.27-0.82$^\circ$, significant at the 3-$\sigma$ level, is found for 5 standards, including the Luminous Blue Variable HD 160529 and all but one of the other B/A-type supergiants (HD 80558, HD 111613, HD 183143 and 55 Cyg), most of which also appear likely to be variable in polarization magnitude ($p$) - there is no preferred orientation for the polarization in these objects, which are all classified as $\alpha$ Cygni variables. Despite this we make six key recommendations for observers - relating to data acquisition, processing and reporting - that will allow them to use these standards to achieve $<$ 0.1$^\circ$ precision in the telescope position angle with similar instrumentation, and allow data sets to be combined more accurately., Comment: 30 pages, 8 figures, 17 tables, 4 appendices; accepted to MNRAS. Companion data at: http://www.mira.org/research/polarimetry/PA

The rising popularity of computational workflows is driven by the need for repetitive and scalable data processing, sharing of processing know-how, and transparent methods. As both combined records of analysis and descriptions of processing steps, workflows should be reproducible, reusable, adaptable, and available. Workflow sharing presents opportunities to reduce unnecessary reinvention, promote reuse, increase access to best practice analyses for non-experts, and increase productivity. In reality, workflows are scattered and difficult to find, in part due to the diversity of available workflow engines and ecosystems, and because workflow sharing is not yet part of research practice. WorkflowHub provides a unified registry for all computational workflows that links to community repositories, and supports both the workflow lifecycle and making workflows findable, accessible, interoperable, and reusable (FAIR). By interoperating with diverse platforms, services, and external registries, WorkflowHub adds value by supporting workflow sharing, explicitly assigning credit, enhancing FAIRness, and promoting workflows as scholarly artefacts. The registry has a global reach, with hundreds of research organisations involved, and more than 700 workflows registered., Comment: 30 pages, 4 figures

Failure mode and effects analysis (FMEA) is a systematic approach to identify and analyse potential failures and their effects in a system or process. The FMEA approach, however, requires domain experts to manually analyse the FMEA model to derive risk-reducing actions that should be applied. In this paper, we provide a formal framework to allow for automatic planning and acting in FMEA models. More specifically, we cast the FMEA model into a Markov decision process which can then be solved by existing solvers. We show that the FMEA approach can not only be used to support medical experts during the modelling process but also to automatically derive optimal therapies for the treatment of patients., Comment: Accepted to the German Journal of Artificial Intelligence

We calculated the structural, electronic and magnetic properties of FeSe within density-functional theory at the generalized gradient approximation level. First, we studied how the bandwidth of the d-bands at the Fermi energy are renormalized by adding simple corrections: Hubbard U, Hunds J and by introducing long-range magnetic orders. We found that introducing either a striped or a staggered dimer antiferromagnetic order brings the bandwidths -- which are starkly overestimated at the generalized gradient approximation level -- closer to those experimentally observed. Second, for the ferromagnetic, the striped, checkerboard and the staggered dimer antiferromagnetic order, we investigate the change in magnetic formation energy with local magnetic moment of Fe at a pressure up to 6 GPa. The bilinear and biquadratic exchange energies are derived from the Heisenberg model and noncollinear first-principles calculations, respectively. We found a non-trivial behavior of the spin-exchange parameters on the magnetization, and we put forward a field-theory model that rationalizes these results in terms of two-dimensional spin and orbital fluctuations. The character of these fluctuations can be either that of a standard density wave or a topological vortex. Topological vortexes can result in mesoscopic magnetization structures.

We present an extension of the quasiparticle self-consistent $GW$ approximation (QS$GW$) [Phys. Rev. B, 76 165106 (2007)] to include vertex corrections in the screened Coulomb interaction $W$. This is achieved by solving the Bethe-Salpeter equation for the polarization matrix at all $k$-points in the Brillouin zone. We refer to this method as QS$G\hat{W}$. QS$GW$ yields a reasonable and consistent description of the electronic structure and optical response, but systematic errors in several properties appear, notably a tendency to overestimate insulating bandgaps, blue-shift plasmon peaks in the imaginary part of the dielectric function, and underestimate the dielectric constant $\epsilon_{\infty}$. A primary objective of this paper is to assess to what extent including ladder diagrams in $W$ ameliorates systematic errors for insulators in the QS$GW$ approximation. For benchmarking we consider about 40 well understood semiconductors, and also examine a variety of less well characterized nonmagnetic systems, six antiferromagnetic oxides, and the ferrimagnet Fe$_3$O$_4$. We find ladders ameliorate shortcomings in QS$GW$ to a remarkable degree in both the one-body Green's function and the dielectric function for a wide range of insulators. New discrepancies with experiment appear, and a key aim of this paper is to establish to what extent the errors are systematic and can be traced to diagrams missing from the theory. One key finding of this work is to establish a relation between the bandgap and the dielectric constant $\epsilon_{\infty}$. Good description of both properties together provides a much more robust benchmark than either alone. We show how this information can be used to improve our understanding of the one-particle spectral properties in materials systems such as SrTiO$_3$ and FeO., Comment: 47 pages, 28 figures

We present a Floquet scheme for the ab-initio calculation of nonlinear optical properties in extended systems. This entails a reformulation of the real-time approach based on the dynamical Berry-phase polarisation [Attaccalite & Gr\"uning, PRB 88, 1-9 (2013)] and retains the advantage of being non-perturbative in the electric field. The proposed method applies to periodically-driven Hamiltonians and makes use of this symmetry to turn a time-dependent problem into a self-consistent time-independent eigenvalue problem. We implemented this Floquet scheme at the independent particle level and compared it with the real-time approach. Our reformulation reproduces real-time-calculated $2^{nd}$ and $3^{rd}$ order susceptibilities for a number of bulk and two-dimensional materials, while reducing the associated computational cost by one or two orders of magnitude.

Min-Nets are inspired by end-stopped cortical cells with units that output the minimum of two learned filters. We insert such Min-units into state-of-the-art deep networks, such as the popular ResNet and DenseNet, and show that the resulting Min-Nets perform better on the Cifar-10 benchmark. Moreover, we show that Min-Nets are more robust against JPEG compression artifacts. We argue that the minimum operation is the simplest way of implementing an AND operation on pairs of filters and that such AND operations introduce a bias that is appropriate given the statistics of natural images.

In contrast to Connectionist Temporal Classification (CTC) approaches, Sequence-To-Sequence (S2S) models for Handwritten Text Recognition (HTR) suffer from errors such as skipped or repeated words which often occur at the end of a sequence. In this paper, to combine the best of both approaches, we propose to use the CTC-Prefix-Score during S2S decoding. Hereby, during beam search, paths that are invalid according to the CTC confidence matrix are penalised. Our network architecture is composed of a Convolutional Neural Network (CNN) as visual backbone, bidirectional Long-Short-Term-Memory-Cells (LSTMs) as encoder, and a decoder which is a Transformer with inserted mutual attention layers. The CTC confidences are computed on the encoder while the Transformer is only used for character-wise S2S decoding. We evaluate this setup on three HTR data sets: IAM, Rimes, and StAZH. On IAM, we achieve a competitive Character Error Rate (CER) of 2.95% when pretraining our model on synthetic data and including a character-based language model for contemporary English. Compared to other state-of-the-art approaches, our model requires about 10-20 times less parameters. Access our shared implementations via this link to GitHub: https://github.com/Planet-AI-GmbH/tfaip-hybrid-ctc-s2s., Comment: 15 pages, 6 tables, 3 figures

Convergence with respect to the size of the k-points sampling-grid of the Brillouin zone is the main bottleneck in the calculation of optical spectra of periodic crystals via the Bethe-Salpeter equation (BSE). We tackle this challenge by proposing a double grid approach to k-sampling compatible with the effective Lanczos-based Haydock iterative solution. Our method relies on a coarse k-grid that drives the computational cost, while a dense k-grid is responsible for capturing excitonic effects, albeit in an approximated way. Importantly, the fine k-grid requires minimal extra computation due to the simplicity of our approach, which also makes the latter straightforward to implement. We performed tests on bulk Si, bulk GaAs and monolayer MoS2, all of which produced spectra in good agreement with data reported elsewhere. This framework has the potential of enabling the calculation of optical spectra in semiconducting systems where the efficiency of the Haydock scheme alone is not enough to achieve a computationally tractable solution of the BSE, e.g., large-scale systems with very stringent k-sampling requirements for achieving convergence., Comment: 31 pages, 7 figures

An increasing number of researchers support reproducibility by including pointers to and descriptions of datasets, software and methods in their publications. However, scientific articles may be ambiguous, incomplete and difficult to process by automated systems. In this paper we introduce RO-Crate, an open, community-driven, and lightweight approach to packaging research artefacts along with their metadata in a machine readable manner. RO-Crate is based on Schema$.$org annotations in JSON-LD, aiming to establish best practices to formally describe metadata in an accessible and practical way for their use in a wide variety of situations. An RO-Crate is a structured archive of all the items that contributed to a research outcome, including their identifiers, provenance, relations and annotations. As a general purpose packaging approach for data and their metadata, RO-Crate is used across multiple areas, including bioinformatics, digital humanities and regulatory sciences. By applying "just enough" Linked Data standards, RO-Crate simplifies the process of making research outputs FAIR while also enhancing research reproducibility. An RO-Crate for this article is available at https://w3id.org/ro/doi/10.5281/zenodo.5146227, Comment: 44 pages. Accepted for Data Science

The optical response functions and band structures of LiCoO$_2$ are studied at different levels of approximation, from density functional theory (DFT) in the generalized gradient approximation (GGA) to quasiparticle self-consistent QS$GW$ (with $G$ for Green's function and $W$ for screened Coulomb interaction) without and with ladder diagrams (QS$G\hat W$) and the Bethe Salpeter Equation (BSE) approach. The QS$GW$ method is found to strongly overestimate the band gap and electron-hole or excitonic effects are found to be important. They lower the quasiparticle gap by only about 11~\% but the lowest energy peaks in absorption are found to be excitonic in nature. The contributions from different band to band transitions and the relation of excitons to band-to-band transitions are analyzed. The excitons are found to be strongly localized. A comparison to experimental data is presented., Comment: 12 pages, 10 figures

We explore the electronic band structure of free standing monolayers of chromium trihalides, CrX\textsubscript{3}{, X= Cl, Br, I}, within an advanced \emph{ab-initio} theoretical approach based in the use of Green's function functionals. We compare the local density approximation with the quasi-particle self-consistent \emph{GW} approximation (QS\emph{GW}) and its self-consistent extension (QS$G\widehat{W}$) by solving the particle-hole ladder Bethe-Salpeter equations to improve the effective interaction \emph{W}. We show that at all levels of theory, the valence band consistently changes shape in the sequence Cl{\textrightarrow}Br{\textrightarrow}I, and the valence band maximum shifts from the M point to the $\Gamma$ point. However, the details of the transition, the one-particle bandgap, and the eigenfunctions change considerably going up the ladder to higher levels of theory. The eigenfunctions become more directional, and at the M point there is a strong anisotropy in the effective mass. Also the dynamic and momentum dependent self energy shows that QS$G\widehat{W}$ adds to the localization of the systems in comparison to the QS\emph{GW} thereby leading to a narrower band and reduced amount of halogens in the valence band manifold.

We present an approach to calculate the electronic structure for a range of materials using the quasiparticle self-consistent GW method with vertex corrections included in the screened Coulomb interaction W. This is achieved by solving the Bethe-Salpeter equation for the polarization matrix at all k-points in the Brillouin zone. We refer to this method as QSGW^. We show that including ladder diagrams in W can greatly reduce the band gap overestimation of RPA-based QSGW. The resultant discrepency of the calculated band gap in this method is then attributed mostly to the fact that electron-phonon contributions to W are neglected; which would allow one to then obtain an estimate for the size of this effect. We present results for a range of systems from simple sp semiconductors to the strongly correlated systems NiO and CoO. Results for systems where the RPA-based QSGW band gap is larger than expected are investigated, and an estimate for the Frolich contribution to the gap is included in a few polar compounds where QSGW can overestimate the gap by as much as 2 eV. The improvement over QSGW for the dielectric constants is also presented, Comment: More up to date version

Functionality of living cells is inherently linked to subunits with dimensions on the nanoscale. In case of osteoblasts the cell surface plays a particularly important role for adhesion and spreading which are crucial properties with regard to bone implants. Here we present a comprehensive characterization of the 3D nanomorphology of living as well as fixed osteoblastic cells using scanning ion conductance microscopy (SICM) which is a nanoprobing method largely avoiding forces. Dynamic ruffles are observed, manifesting themselves in characteristic membrane protrusions. They contribute to the overall surface corrugation which we systematically study by introducing the relative 3D excess area as a function of projected adhesion area. A clear anticorrelation is found upon analysis of ~40 different cells on glass as well as on amine covered surfaces. At the rim of lamellipodia characteristic edge heights between 100 nm and ~300 nm are observed. Power spectral densities of membrane fluctuations show frequency-dependent decay exponents in excess of -2 on living osteoblasts. We discuss the capability of apical membrane features and fluctuation dynamics in aiding assessment of adhesion and migration properties on a single-cell basis., Comment: 27 pages, 11 figures

We introduce Feature-Product networks (FP-nets) as a novel deep-network architecture based on a new building block inspired by principles of biological vision. For each input feature map, a so-called FP-block learns two different filters, the outputs of which are then multiplied. Such FP-blocks are inspired by models of end-stopped neurons, which are common in cortical areas V1 and especially in V2. Convolutional neural networks can be transformed into parameter-efficient FP-nets by substituting conventional blocks of regular convolutions with FP-blocks. In this way, we create several novel FP-nets based on state-of-the-art networks and evaluate them on the Cifar-10 and ImageNet challenges. We show that the use of FP-blocks reduces the number of parameters significantly without decreasing generalization capability. Since so far heuristics and search algorithms have been used to find more efficient networks, it seems remarkable that we can obtain even more efficient networks based on a novel bio-inspired design principle.

Experimental studies support the notion of spike-based neuronal information processing in the brain, with neural circuits exhibiting a wide range of temporally-based coding strategies to rapidly and efficiently represent sensory stimuli. Accordingly, it would be desirable to apply spike-based computation to tackling real-world challenges, and in particular transferring such theory to neuromorphic systems for low-power embedded applications. Motivated by this, we propose a new supervised learning method that can train multilayer spiking neural networks to solve classification problems based on a rapid, first-to-spike decoding strategy. The proposed learning rule supports multiple spikes fired by stochastic hidden neurons, and yet is stable by relying on first-spike responses generated by a deterministic output layer. In addition to this, we also explore several distinct, spike-based encoding strategies in order to form compact representations of presented input data. We demonstrate the classification performance of the learning rule as applied to several benchmark datasets, including MNIST. The learning rule is capable of generalising from the data, and is successful even when used with constrained network architectures containing few input and hidden layer neurons. Furthermore, we highlight a novel encoding strategy, termed `scanline encoding', that can transform image data into compact spatiotemporal patterns for subsequent network processing. Designing constrained, but optimised, network structures and performing input dimensionality reduction has strong implications for neuromorphic applications., Comment: 41 pages, 13 figures

The temperature renormalization of the bulk band structure of a topological crystalline insulator, SnTe, is calculated using first principles methods. We explicitly include the effect of thermal-expansion-induced modification of electronic states and their band inversion on electron-phonon interaction. We show that the direct gap decreases with temperature, as both thermal expansion and electron-phonon interaction drive SnTe towards the phase transition to a topologically trivial phase as temperature increases. The band gap renormalization due to electron-phonon interaction exhibits a non-linear dependence on temperature as the material approaches the phase transition, while the lifetimes of the conduction band states near the band edge show a non-monotonic behavior with temperature. These effects should have important implications on bulk electronic and thermoelectric transport in SnTe and other topological insulators., Comment: 10 pages, 8 figures. Accepted for publication in Phys. Rev. B on June 8, 2020

PbTe is a leading thermoelectric material at intermediate temperatures, largely thanks to its low lattice thermal conductivity. However, its efficiency is too low to compete with other forms of power generation. This efficiency can be effectively enhanced by designing nanostructures capable of scattering phonons over a wide range of length scales to reduce the lattice thermal conductivity. The presence of grain boundaries can reduce the thermal conductivity to $\sim 0.5$ Wm$^{-1}$K$^{-1}$ for small vacancy concentrations and grain sizes. However, grains anneal at finite temperature, and equilibrium and metastable grain size distributions determine the extent of the reduction in thermal conductivity. In the present work, we propose a phase-field model informed by molecular dynamics simulations to study the annealing process in PbTe and how it is affected by the presence of grain boundaries and voids. We find that the thermal conductivity of PbTe is reduced by up to 35\% in the porous material at low temperatures. We observe that a phase transition at a finite density of voids governs the kinetics of impeding grain growth by Zener pinning.

We calculate the electronic structure of the narrow gap semiconductors PbTe, SnTe and GeTe in the cubic phase using density functional theory (DFT) and the $G_0W_0$ method. Within DFT, we show that the band ordering obtained with a conventional semilocal exchange-correlation approximation is correct for SnTe and GeTe but wrong for PbTe. The correct band ordering at the high-symmetry point L is recovered adding $G_0W_0$ quasiparticle corrections. However, one-shot $G_0W_0$ produces artifacts in the band structure due to the wrong orbital character of the DFT single-particle states at the band edges close to L. We show that in order to correct these artifacts it is enough to consider the off-diagonal elements of the $G_0W_0$ self-energy corresponding to these states. We also investigate the pressure dependence of the band gap for these materials and the possibility of a transition from a trivial to a non-trivial topology of the band structure. For PbTe, we predict the band crossover and topological transition to occur at around 4.8 GPa. For GeTe, we estimate the topological transition to occur at 1.9 GPa in the constrained cubic phase, a pressure lower than the one of the structural phase transition from rombohedral to cubic. SnTe is a crystalline topological insulator at ambient pressure, and the transition into a trivial topology would take place under a volume expansion of approximately $10\%$., Comment: 8 pages, 4 figures

Spiking neural networks (SNNs) are distributed trainable systems whose computing elements, or neurons, are characterized by internal analog dynamics and by digital and sparse synaptic communications. The sparsity of the synaptic spiking inputs and the corresponding event-driven nature of neural processing can be leveraged by energy-efficient hardware implementations, which can offer significant energy reductions as compared to conventional artificial neural networks (ANNs). The design of training algorithms lags behind the hardware implementations. Most existing training algorithms for SNNs have been designed either for biological plausibility or through conversion from pretrained ANNs via rate encoding. This article provides an introduction to SNNs by focusing on a probabilistic signal processing methodology that enables the direct derivation of learning rules by leveraging the unique time-encoding capabilities of SNNs. We adopt discrete-time probabilistic models for networked spiking neurons and derive supervised and unsupervised learning rules from first principles via variational inference. Examples and open research problems are also provided., Comment: Published in IEEE Signal Processing Magazine, Vol. 36, No. 6, pp. 64-77 (subsumes arXiv:1812.03929), Author's Accepted Manuscript

Measuring the performance of text recognition and text line detection engines is an important step to objectively compare systems and their configuration. There exist well-established measures for both tasks separately. However, there is no sophisticated evaluation scheme to measure the quality of a combined text line detection and text recognition system. The F-measure on word level is a well-known methodology, which is sometimes used in this context. Nevertheless, it does not take into account the alignment of hypothesis and ground truth text and can lead to deceptive results. Since users of automatic information retrieval pipelines in the context of text recognition are mainly interested in the end-to-end performance of a given system, there is a strong need for such a measure. Hence, we present a measure to evaluate the quality of an end-to-end text recognition system. The basis for this measure is the well established and widely used character error rate, which is limited -- in its original form -- to aligned hypothesis and ground truth texts. The proposed measure is flexible in a way that it can be configured to penalize different reading orders between the hypothesis and ground truth and can take into account the geometric position of the text lines. Additionally, it can ignore over- and under- segmentation of text lines. With these parameters it is possible to get a measure fitting best to its own needs., Comment: to appear in proceeding at ICDAR 2019

Encoder-decoder models have become an effective approach for sequence learning tasks like machine translation, image captioning and speech recognition, but have yet to show competitive results for handwritten text recognition. To this end, we propose an attention-based sequence-to-sequence model. It combines a convolutional neural network as a generic feature extractor with a recurrent neural network to encode both the visual information, as well as the temporal context between characters in the input image, and uses a separate recurrent neural network to decode the actual character sequence. We make experimental comparisons between various attention mechanisms and positional encodings, in order to find an appropriate alignment between the input and output sequence. The model can be trained end-to-end and the optional integration of a hybrid loss allows the encoder to retain an interpretable and usable output, if desired. We achieve competitive results on the IAM and ICFHR2016 READ data sets compared to the state-of-the-art without the use of a language model, and we significantly improve over any recent sequence-to-sequence approaches., Comment: 8 pages, 1 figure, 8 tables

Second Harmonic Generation (SHG) of single-layer monochalcogenides, such as GaSe and InSe, has been recently reported [2D Mater. 5 (2018) 025019; J. Am. Chem. Soc. 2015, 137, 79947997] to be extremely strong with respect to bulk and multilayer forms. To clarify the origin of this strong SHG signal, we perform first-principles real-time simulations of linear and non-linear optical properties of these two-dimensional semiconducting materials. The simulations, based on ab-initio many-body theory, accurately treat the electron-hole correlation and capture excitonic effects that are deemed important to correctly predict the optical properties of such systems. We find indeed that, as observed for other 2D systems, the SHG intensity is redistributed at excitonic resonances. The obtained theoretical SHG intensity is an order of magnitude smaller than that reported at the experimental level. This result is in substantial agreement with previously published simulations which neglected the electron-hole correlation, demonstrating that many-body interactions are not at the origin of the strong SHG measured. We then show that the experimental data can be reconciled with the theoretical prediction when a single layer model, rather than a bulk one, is used to extract the SHG coefficient from the experimental data., Comment: 8 pages, 4 figures

yambo is an open source project aimed at studying excited state properties of condensed matter systems from first principles using many-body methods. As input, yambo requires ground state electronic structure data as computed by density functional theory codes such as quantum-espresso and abinit. yambo's capabilities include the calculation of linear response quantities (both independent-particle and including electron-hole interactions), quasi-particle corrections based on the GW formalism, optical absorption, and other spectroscopic quantities. Here we describe recent developments ranging from the inclusion of important but oft-neglected physical effects such as electron-phonon interactions to the implementation of a real-time propagation scheme for simulating linear and non-linear optical properties. Improvements to numerical algorithms and the user interface are outlined. Particular emphasis is given to the new and efficient parallel structure that makes it possible to exploit modern high performance computing architectures. Finally, we demonstrate the possibility to automate workflows by interfacing with the yambopy and AiiDA software tools.

VO$_{2}$ is a model material system which exhibits a metal to insulator transition at 67$^\circ$C. This holds potential for future ultrafast switching in memory devices, but typically requires a purely electronic process to avoid the slow lattice response. The role of lattice vibrations is thus important, but it is not well understood and it has been a long-standing source of controversy. We use a combination of ultrafast spectroscopy and ab initio quantum calculations to unveil the mechanism responsible for the transition. We identify an atypical Peierls vibrational mode which acts as a trigger for the transition. This rules out the long standing paradigm of a purely electronic Mott transition in VO$_{2}$; however, we found a new electron-phonon pathway for a purely reversible electronic transition in a true bi-stable fashion under specific conditions. This transition is very atypical, as it involves purely charge-like excitations and requires only small nuclear displacement. Our findings will prompt the design of future ultrafast electro-resistive non-volatile memory devices., Comment: 13 pages, 3 figures

Spiking Neural Networks (SNNs) are distributed trainable systems whose computing elements, or neurons, are characterized by internal analog dynamics and by digital and sparse synaptic communications. The sparsity of the synaptic spiking inputs and the corresponding event-driven nature of neural processing can be leveraged by hardware implementations that have demonstrated significant energy reductions as compared to conventional Artificial Neural Networks (ANNs). Most existing training algorithms for SNNs have been designed either for biological plausibility or through conversion from pre-trained ANNs via rate encoding. This paper aims at providing an introduction to SNNs by focusing on a probabilistic signal processing methodology that enables the direct derivation of learning rules leveraging the unique time encoding capabilities of SNNs. To this end, the paper adopts discrete-time probabilistic models for networked spiking neurons, and it derives supervised and unsupervised learning rules from first principles by using variational inference. Examples and open research problems are also provided., Comment: This article is now superseded by arXiv:1910.01059. To appear on IEEE Signal Processing Magazine

We present a recognition and retrieval system for the ICDAR2017 Competition on Information Extraction in Historical Handwritten Records which successfully infers person names and other data from marriage records. The system extracts information from the line images with a high accuracy and outperforms the baseline. The optical model is based on Neural Networks. To infer the desired information, regular expressions are used to describe the set of feasible words sequences.

First principles electronic structure calculations are carried out to investigate the band alignments of tensile strained (001) Ge interfaced with (001) In$_{x}$Al$_{1-x}$As. The sensitivities of band offsets to interfacial structure, interfacial stoichiometry, and substrate stoichiometry, are investigated. Large qualitative variations of the valence and conduction band offsets are observed, including changes of the band offset type, indicating the importance of local structural variations of the interface for band offsets in real samples. Our results explain recent measurements of band offsets derived from XPS core level spectra in terms of As atoms penetrating through the first few monolayers of the Ge film. Analogous studies are carried out for the diffusion of other species across the interface, and in general the band offsets vary approximately linearly with diffusion depth relative to the values for pristine "sharp" interfaces, where the sign of the linear variation depends on the diffusing species. This large sensitivity of the band alignments to interface details indicates potential routes to chemically control the band offset of this group IV/III-V interface by tuning the stoichiometry of the substrate surface that the thin film is grown on., Comment: 12 pages, 10 figures

We calculate the two-photon absorption in bulk and single layer hexagonal boron nitride (hBN) both by an ab-initio real-time Bethe-Salpeter approach and by a the real-space solution of the excitonic problem in tight-binding formalism. The two-photon absorption obeys different selection rules from those governing linear optics and therefore provides complementary information on the electronic excitations of hBN. Combining the results from the simulations with a symmetry analysis we show that two-photon absorption is able to probe the lowest energy $1s$ states in the single layer hBN and the lowest dark degenerate dark states of bulk hBN. This deviation from the "usual" selection rules based on the continuous hydrogenic model is explained within a simple model that accounts for the crystalline symmetry. The same model can be applied to other two-dimensional materials with the same point-group symmetry, such as the transition metal chalcogenides. We also discuss the selection rules related to the inversion symmetry of the bulk layer stacking., Comment: 13 pages, 4 figures

For many 2D materials, optical and Raman spectra are richly structured, and convey information on a range of parameters including nanosheet size and defect content. By contrast, the equivalent spectra for h-BN are relatively simple, with both the absorption and Raman spectra consisting of a single feature each, disclosing relatively little information. Here, the ability to size-select liquid-exfoliated h-BN nanosheets has allowed us to comprehensively study the dependence of h-BN optical spectra on nanosheet dimensions. We find the optical extinction coefficient spectrum to vary systematically with nanosheet lateral size due to the presence of light scattering. Conversely, once light scattering has been decoupled to give the optical absorbance spectra, we find the size dependence to be mostly removed save for a weak but well-defined variation in energy of peak absorbance with nanosheet thickness. This finding is corroborated by our ab initio GW and Bethe-Salpeter equation calculations, which include electron correlations and quasiparticle self-consistency (QSGW). In addition, while we find the position of the sole h-BN Raman line to be invariant with nanosheet dimensions, the linewidth appears to vary weakly with nanosheet thickness. These size-dependent spectroscopic properties can be used as metrics to estimate nanosheet thickness from spectroscopic data., Comment: Accepted in Chemistry Materials (In press)

It is known that a geodesic Y in an abstract reflection space X in the sense of Loos, without any assumption of differential structure, canonically admits an action of a 1-parameter subgroup of the group of transvections of X. In this article, we prove an analog of this result stating that, if X contains an embedded hyperbolic plane H, then this yields a canonical action of a subgroup of the transvection group of X isomorphic to a perfect central extension of PSL(2,R). This result can be further extended to arbitrary Riemannian symmetric spaces of non-compact type embedded in X and can be used to prove that a Riemannian symmetric space and, more generally, the Kac-Moody symmetric space G/K for an algebraically simply connected two-spherical Kac-Moody group G satisfies a universal property similar to the universal property that the group G satisfies itself.

This work presents a two-stage text line detection method for historical documents. Each detected text line is represented by its baseline. In a first stage, a deep neural network called ARU-Net labels pixels to belong to one of the three classes: baseline, separator or other. The separator class marks beginning and end of each text line. The ARU-Net is trainable from scratch with manageably few manually annotated example images (less than 50). This is achieved by utilizing data augmentation strategies. The network predictions are used as input for the second stage which performs a bottom-up clustering to build baselines. The developed method is capable of handling complex layouts as well as curved and arbitrarily oriented text lines. It substantially outperforms current state-of-the-art approaches. For example, for the complex track of the cBAD: ICDAR2017 Competition on Baseline Detection the F-value is increased from 0.859 to 0.922. The framework to train and run the ARU-Net is open source., Comment: to be published in IJDAR

We present an approach to calculate the optical absorption spectra that combines the quasiparticle self-consistent GW method [Phys. Rev. B, 76 165106 (2007)] for the electronic structure with the solution of the ladder approximation to the Bethe-Salpeter equation for the macroscopic dielectric function. The solution of the Bethe-Salpeter equation has been implemented within an all-electron framework, using a linear muffin-tin orbital basis set, with the contribution from the non-local self-energy to the transition dipole moments (in the optical limit) evaluated explicitly. This approach addresses those systems whose electronic structure is poorly described within the standard perturbative GW approaches with as a starting point density-functional theory calculations. The merits of this approach have been exemplified by calculating optical absorption spectra of a strongly correlated transition metal oxide, NiO, and a narrow gap semiconductor, Ge. In both cases, the calculated spectrum is in good agreement with the experiment. It is also shown that for systems whose electronic structure is well-described within the standard perturbative GW, such as Si, LiF and h-BN, the performance of the present approach is in general comparable to the standard GW plus Bethe-Salpeter equation. It is argued that both vertex corrections to the electronic screening and the electron-phonon interaction are responsible for the observed systematic overestimation of the fundamental bandgap and spectrum onset., Comment: 16 pages, 9 figures