Search

Abstract Reservoir computing advances the intriguing idea that a nonlinear recurrent neural circuit—the reservoir—can encode spatio-temporal input signals to enable efficient ways to perform tasks like classification or regression. However, recently the idea of a monolithic reservoir network that simultaneously buffers input signals and expands them into nonlinear features has been challenged. A representation scheme in which memory buffer and expansion into higher-order polynomial features can be configured separately has been shown to significantly outperform traditional reservoir computing in prediction of multivariate time-series. Here we propose a configurable neuromorphic representation scheme that provides competitive performance on prediction, but with significantly better scaling properties than directly materializing higher-order features as in prior work. Our approach combines the use of randomized representations from traditional reservoir computing with mathematical principles for approximating polynomial kernels via such representations. While the memory buffer can be realized with standard reservoir networks, computing higher-order features requires networks of ‘Sigma-Pi’ neurons, i.e., neurons that enable both summation as well as multiplication of inputs. Finally, we provide an implementation of the memory buffer and Sigma-Pi networks on Loihi 2, an existing neuromorphic hardware platform.

Ti-6.5Al-2Zr-1Mo-1V/TiB metal matrix composites with 3 wt.% of TiB2 were obtained using vacuum arc melting and spark plasma sintering methods and compared with an unreinforced Ti-6.5Al-2Zr-1Mo-1V alloy. The microstructures of the unreinforced Ti6.5Al-2Zr-1Mo-1V alloy in the as-cast and as-sintered conditions were quite typical and consisted of colonies of α-lamellae embedded in the β matrix. The microstructure of the as-cast Ti-6.5Al-2Zr-1Mo-1V/TiB composite composed of TiB fibers randomly distributed within the two-phase α/β matrix, while the as-sintered composite had a network-like microstructure, in which areas of the two-phase α/β matrix were delineated by walls of TiB fibers. At room temperature, the yield strength of the as-cast and as-sintered Ti-6.5Al-2Zr-1Mo-1V alloy were 800 and 915 MPa, respectively, with a plasticity of 18% in both conditions. The addition of TiB fibers contributed to a ~40 and 50% strength increment, with values of 1100 and 1370 MPa for the as-cast and as-sintered composites, respectively. In the as-sintered composite, the strengthening effect reduced at 400 °C and almost disappeared at elevated temperatures of 800–950 °C. The as-cast composite showed much higher strength during warm and hot deformation—at 800–950 °C, the yield strength of the as-cast composite was 50% higher compared to the Ti-6.5Al-2Zr-1Mo-1V unreinforced alloy. A higher rate and degree of globularization were established for the as-cast composite compared to the unreinforced alloy. For the as-sintered composite, a noticeably lower rate and degree of globularization was shown. During hot compression of the as-cast composite, TiB fibers reoriented towards the metal flow direction, while the network microstructure formed in the as-sintered composite transformed into clusters of borides unevenly distributed within the matrix. Based on the obtained results, the apparent activation energy of plastic deformation was calculated, and the operating deformation mechanisms were discussed both for the as-cast and as-sintered composites. The Arrhenius flow stress model and the dynamic material model were used to evaluate the deformation behavior of composites beyond the experimentally studied temperatures and strain rates.

Abstract A prominent approach to solving combinatorial optimization problems on parallel hardware is Ising machines, i.e., hardware implementations of networks of interacting binary spin variables. Most Ising machines leverage second-order interactions although important classes of optimization problems, such as satisfiability problems, map more seamlessly to Ising networks with higher-order interactions. Here, we demonstrate that higher-order Ising machines can solve satisfiability problems more resource-efficiently in terms of the number of spin variables and their connections when compared to traditional second-order Ising machines. Further, our results show on a benchmark dataset of Boolean k-satisfiability problems that higher-order Ising machines implemented with coupled oscillators rapidly find solutions that are better than second-order Ising machines, thus, improving the current state-of-the-art for Ising machines.

Aim: The aim was to define the impact of comorbidities, specifically hypertension as one of the most common chronic diseases, on the outcome and length of stay for COVID-19 patients. Methods: The cross-sectional study, carried out from October to December 2021, included 129 hospitalised COVID-19 patients who presented to the Emergency Department and were hospitalised and treated in the COVID ward in the General Hospital Našice. All patients tested positive for COVID-19 with a polymerase chain reaction (PCR) test. Clinical parameters were also recorded and they included demographic factors, comorbidities, type of antihypertensive therapy, new-onset hypertension, length of stay and the overall outcome. Results: The most common comorbidity was hypertension (86, 66.7%). Hypertension was associated with women (P = 0.03) and age over 65 years (P < 0.001). Length of stay was longer for patients with hypertension (P = 0.04) and/or diabetes mellitus (P = 0.04). Higher mortality was associated with age over 65 years (P < 0.001) and a higher number of comorbidities (P = 0.01). New-onset hypertension was recorded in three patients. There was no significant difference in the outcome in relation to antihypertensive therapy. Conclusion: Hypertension is the most common comorbidity in hospitalised COVID-19 patients. Although treated hypertension did not have a negative impact on the outcome, other potential risk factors, including a higher number of comorbidities and older age, are associated with mortality in COVID-19 patients.

Data augmentation is an important method for evaluating the robustness of and enhancing the diversity of training data for natural language processing (NLP) models. In this paper, we present NL-Augmenter, a new participatory Python-based natural language (NL) augmentation framework which supports the creation of transformations (modifications to the data) and filters (data splits according to specific features). We describe the framework and an initial set of 117 transformations and 23 filters for a variety of NL tasks annotated with noisy descriptive tags. The transformations incorporate noise, intentional and accidental human mistakes, socio-linguistic variation, semantically-valid style, syntax changes, as well as artificial constructs that are unambiguous to humans. We demonstrate the efficacy of NL-Augmenter by using its transformations to analyze the robustness of popular language models. We find different models to be differently challenged on different tasks, with quasi-systematic score decreases. The infrastructure, datacards, and robustness evaluation results are publicly available on GitHub for the benefit of researchers working on paraphrase generation, robustness analysis, and low-resource NLP. El aumento de datos es un método importante para evaluar la solidez y mejorar la diversidad del entrenamiento datos para modelos de procesamiento de lenguaje natural (NLP). इस लेख में, हम एनएल-ऑगमेंटर का प्रस्ताव करते हैं - एक नया भागी- दारी पूर्वक, पायथन में बनाया गया, लैंग्वेज (एनएल) ऑग्मेंटेशन फ्रेमवर्क जो ट्रांसफॉर्मेशन (डेटा में बदलाव करना) और फीलटर (फीचर्स के अनुसार डेटा का भाग करना) के नीरमान का समर्थन करता है।. 我们描述了NL-Augmenter框架及其初步包含的117种转换和23个过滤器，并 大致标注分类了一系列可适配的自然语言任务. این دگرگونی ها شامل نویز، اشتباهات عمدی و تصادفی انسانی، تنوع اجتماعی-زبانی، سبک معنایی معتبر، تغییرات نحوی و همچنین ساختارهای مصنوعی است که برای انسان ها مبهم است. NL-Augmenterpa allin kaynintam qawachiyku, tikrakuyninku- nata servichikuspayku, chaywanmi qawariyku modelos de lenguaje popular nisqapa allin takyasqa kayninta. Kami menemukan model yang berbeda ditantang secara berbeda pada tugas yang berbeda, dengan penurunan skor kuasi-sistematis. Infrastruktur, kartu data, dan hasil evaluasi ketahanan dipublikasikan tersedia secara gratis di GitHub untuk kepentingan para peneliti yang mengerjakan pembuatan parafrase, analisis ketahanan, dan NLP sumber daya rendah.

Operations on high-dimensional, fixed-width vectors can be used to distribute information from several vectors over a single vector of the same width. For example, a set of key-value pairs can be encoded into a single vector with multiplication and addition of the corresponding key and value vectors: the keys are bound to their values with component-wise multiplication, and the key-value pairs are combined into a single superposition vector with component-wise addition. The superposition vector is, thus, a memory which can then be queried for the value of any of the keys, but the result of the query is approximate. The exact vector is retrieved from a codebook (a.k.a. item memory), which contains vectors defined in the system. To perform these operations, the item memory vectors and the superposition vector must be the same width. Increasing the capacity of the memory requires increasing the width of the superposition and item memory vectors. In this article, we demonstrate that in a regime where many (e.g., 1,000 or more) key-value pairs are stored, an associative memory which maps key vectors to value vectors requires less memory and less computing to obtain the same reliability of storage as a superposition vector. These advantages are obtained because the number of storage locations in an associate memory can be increased without increasing the width of the vectors in the item memory. An associative memory would not replace a superposition vector as a medium of storage, but could augment it, because data recalled from an associative memory could be used in algorithms that use a superposition vector. This would be analogous to how human working memory (which stores about seven items) uses information recalled from long-term memory (which is much larger than the working memory). We demonstrate the advantages of an associative memory experimentally using the storage of large finite-state automata, which could model the storage and recall of state-dependent behavior by brains.

Russia remains one of the areas most affected by HIV in Eastern Europe and Central Asia. The aim of this study was to analyze HIV infection indicators and study trends in Russia using data from the Federal Statistic Form No. 61 “Information about HIV infection”. HIV incidence, prevalence, HIV testing and mortality rates (from 2011 to 2022), and treatment success rates (from 2016 to 2022) were analyzed. These indicators were compared across different federal districts (FDs) of Russia. The findings revealed a significant downward trend in HIV incidence, while a significant upward trend was observed for HIV prevalence. The mortality rate has stabilized since 2018. The coverage of HIV testing and antiretroviral therapy increased over time. The number of people living with HIV-1 (PLWH) with a suppressed viral load in Russia as a whole varied between 72% and 77% during the years under observation. The Siberian and Ural federal districts recorded the highest HIV incidence, while the North Caucasian FD reported the lowest. An increase in HIV testing coverage was observed across all FDs. This comprehensive evaluation of HIV infection indicators within the regional context contributes to the timely implementation of measures aimed at preventing the spread of HIV.

The United States is currently affected by widespread hepatitis A virus (HAV) outbreaks. We investigated HAV incidence rates among source plasma donors in the United States since 2016. Serial donations from HAV-positive frequent donors were analyzed for common biologic markers to obtain a detailed picture of the course of infection. We found a considerable increase in incidence rates with shifting outbreak hotspots over time. Although individual biomarker profiles were highly variable, HAV RNA typically had a high peak and a biphasic decrease and often remained detectable for several months. One donor had a biomarker pattern indicative of previous exposure. Our findings show that current HAV outbreaks have been spilling over into the plasma donor population. The detailed results presented improve our comprehension of HAV infection and related public health aspects. In addition, the capture of full RNA curves enables estimation of HAV doubling time.

An experimental study of post-dryout heat transfer with the coolant R-134a was performed in a vertical round tube with upward flow. The experiments were conducted in an uniformly heated tube with an inside diameter of 10 mm and a heated length of 3100 mm, at pressures of 11.1 bar, 16 bar and 28 bar, with varying mass flux in the range of 300–1500 kg/m2s.The results show that for higher mass fluxes and higher heat fluxes the wall temperature decreases after reaching a peak close to the critical vapor quality, while at lower mass fluxes the wall temperature keeps increasing after dryout. The phenomenon described last is shifted to even lower mass fluxes with rising pressure. Furthermore, it was found that at increasing pressure as well as at increasing mass flux the dryout location is shifted to lower equilibrium vapor qualities.Based on the experimental results and existing prediction methods, a theoretical model was developed to describe the heat transfer in the post-dryout area. It divides the post-dryout area into a developing and a developed region. A correction factor was introduced to model the droplet distribution in the cross-section of the fully developed region. A comparison with selected post-dryout heat transfer models from the literature for the experimental parameter range in the present study shows a good applicability of the new model.

As it happens with any heretical doctrine that takes part of the Christian truth for a whole, the further it goes from the fullness of the truth, the more it parasitizes on parts of that truth. At present time it is impossible to ignore the increased activity of different sects and the fact that their followers grow in numbers. The activities of these organizations are destructive. Rejecting the authenticity of gospel teaching, they present a serious challenge for the Christians of our time. The most common totalitarian sect of today is the Jehovah's Witnesses, a religious organization founded by Charles Russell in the United States at the end of the nineteenth century. His organization received its present name in 1931. Like other contemporary religious organizations that are not affiliated with the Church of Christ, Jehovah's Witnesses also respond very painfully to the question of the historical connection with the apostolic times and the date of foundation of the organization. The followers of this sect call themselves Christians and are very active in spreading their special teachings around the world. The wording of the Jehovah's Witness doctrine is solely the prerogative of the Governing Body located in Brooklyn. Only this authority has the power to make the Bible’s interpretation normative and binding on all followers. The doctrine can, over time, undergo certain transformations, which is due to the revelation of God, which can only be accepted by the higher leadership. Jehovah's Witnesses practice control of members of the organization, isolation from external relations with society, psychological influence and pressure on dissenters, rigid conspiracy and discipline. The closure from the outside world, which has repeatedly provoked family conflicts, is the cause of sharp criticism of the practices and structure of Jehovah’s Witnesses, as well as accusations of sectarianism and even of a totalitarian approach. This article outlines the major milestones in the history of the formation of the Jehova’s Witnesses, and analyzes how much the Christian principles of the Watchtower Society proposed to their followers and propagated through their missionaries are consistent with Christian teaching.

Autonomic function during deep breathing (DB) is normally scored based on the assumption that the heart rate is synchronized with the breathing. We have observed individuals with subtle arrhythmias during DB, where an autonomic function cannot be evaluated. This paper presents a novel method for analyzing cardiorespiratory synchronization: feature-based analysis of the similarity between heart rate and respiration using the principles of hyper-dimensional computing. Heart rate and respiration signals were modeled using Fourier series analysis. Three feature variables were derived and mapped to binary vectors in a high-dimensional space. Using both synthesized data and recordings from patients/healthy subjects, the similarity between the feature vectors was assessed using Hamming distance (high-dimensional space), Euclidean distance (original space), and with a coherence-based index. Methods were evaluated via the classification of the similarity indices into three groups. The distance-based methods achieved good separation of signals into classes with different degrees of cardiorespiratory synchronization, also providing identification of patients with low cardiorespiratory synchronization but high values of conventional DB scores. Moreover, binary high-dimensional vectors allowed an additional analysis of the obtained Hamming distance. Feature-based similarity analysis using hyper-dimensional computing is capable of identifying signals with low cardiorespiratory synchronization during DB due to arrhythmias. Vector-based similarity analysis could be applied to other types of feature variables than based on spectral analysis. The proposed methods for robustly assessing cardiorespiratory synchronization during DB facilitate the identification of individuals where the evaluation of the autonomic function is problematic or even impossible, thus increasing the correctness of the conventional DB scores.

The avidity index (AI) of IgG to the RBD of SARS-CoV-2 was determined for 71 patients with a mild (outpatient) course of COVID-19, including 39 primarily and 36 secondarily reinfected, and 92 patients with a severe (hospital) course of COVID-19, including 82 primarily and 10 secondarily infected. The AI was shown to correlate with the severity of repeated disease. In the group of outpatients with a mild course, the reinfected patients had significantly higher median AIs than those with primary infections (82.3% vs. 37.1%, p < 0.0001). At the same time, in patients with a severe course of COVID-19, reinfected patients still had low-avidity antibodies (median AI of 28.4% vs. 25% in the primarily infected, difference not significant, p > 0.05). This suggests that the presence of low-avidity IgG to RBD during reinfection is a negative prognostic factor, in which a patient’s risk of developing COVID-19 in a severe form is significantly increased. Thus, patients with IgG of low avidity (AI ≤ 40%) had an 89 ± 20.5% chance of a severe course of recurrent COVID-19, whereas the detection of high-avidity antibodies (AI ≥ 50%) gave a probability of 94 ± 7.9% for a mild course of recurrent disease (p < 0.05).

This paper presents an approach for distributed fault isolation in a generic system of systems. The proposed approach is based on the principles of hyperdimensional computing. In particular, the recently proposed method called Holographic Graph Neuron is used. We present a distributed version of Holographic Graph Neuron and evaluate its performance on the problem of fault isolation in a complex power plant model. Compared to conventional machine learning methods applied in the context of the same scenario the proposed approach shows comparable performance while being distributed and requiring simple binary operations, which allow for a fast and efficient implementation in hardware.

Abstract We analyze exotic matter representations that arise on singular seven-brane configurations in F-theory. We develop a general framework for analyzing such representations, and work out explicit descriptions for models with matter in the 2-index and 3-index symmetric representations of SU(N) and SU(2) respectively, associated with double and triple point singularities in the seven-brane locus. These matter representations are associated with Weierstrass models whose discriminants vanish to high order thanks to nontrivial cancellations possible only in the presence of a non-UFD algebraic structure. This structure can be described using the normalization of the ring of intrinsic local functions on a singular divisor. We consider the connection between geometric constraints on singular curves and corresponding constraints on the low-energy spectrum of 6D theories, identifying some new examples of apparent “swampland” theories that cannot be realized in F-theory but have no apparent low-energy inconsistency.

Discrete gauge groups naturally arise in F-theory compactifications on genus-one fibered Calabi–Yau manifolds. Such geometries appear in families that are parameterized by the Tate–Shafarevich group of the genus-one fibration. While the F-theory compactification on any element of this family gives rise to the same physics, the corresponding M-theory compactifications on these geometries differ and are obtained by a fluxed circle reduction of the former. In this note, we focus on an element of order three in the Tate–Shafarevich group of the general cubic. We discuss how the different M-theory vacua and the associated discrete gauge groups can be obtained by Higgsing of a pair of five-dimensional U(1) symmetries. The Higgs fields arise from vanishing cycles in I2-fibers that appear at certain codimension two loci in the base. We explicitly identify all three curves that give rise to the corresponding Higgs fields. In this analysis the investigation of different resolved phases of the underlying geometry plays a crucial rôle.

As the customer relationship management process comes to play an increasingly important role in business success, a number of authors are attempting to evaluate the impact of various CRM process components on the quality of company interaction with customers and, ultimately, on company performance. This paper explores the impact of CRM systems on the quality of companies’ customer understanding. This understanding is measured in the context of an international pharmaceutical company in the Russian market. The field research is based on quantitative data from online questionnaires and telephone interviews. The sample consists of 64 company representatives and 217 ophthalmologists. The authors developed and tested a model of physician loyalty drivers and studied employees’ perceptions of the CRM system. The findings of this paper demonstrate that, despite the fact that a CRM system is actively used and perceived as a crucial part of the customer relationship management process within the company, understanding of key customer loyalty drivers needs to be significantly improved. The paper contributes to existing theory by evaluating the link between the use of CRM applications and customer relationship performance as well as by developing a physician prescription loyalty influencers framework in the context of the Russian pharmaceutical market. This research could be used by other pharmaceutical companies as well in order to understand the influence of their CRM applications on customer loyalty and also to identify the drivers of physicians’ prescriptions.

We study the bound state spectrum and the conditions for entering a supercritical regime in graphene with strong intrinsic and Rashba spin-orbit interactions within the topological insulator phase. Explicit results are provided for a disk-shaped potential well and for the Coulomb center problem.

Detecting activities of daily living (ADL) allows for rich inference about user behavior, which can be of use in the care of for example, elderly people, chronic diseases, and psychological conditions. This paper proposes a domain knowledge-based solution for detecting 24 different ADLs in the UJA dataset. The solution is inspired by a Finite State Machine and performs activity recognition unobtrusively using only binary sensors. Each day in the dataset is segmented into: morning, day, evening in order to facilitate the inference from the sensors. The model performs the ADL recognition in two steps. The first step is to detect the sequence of activities in a given event stream of binary sensors, and the second step is to assign a starting and ending times for each of detected activities. Our proposed model achieved an accuracy of 81.3% using only a very small amount of operations, making it an interesting approach for resource-constrained devices that are common in smart environments. It should be noted, however, that the model can end up in faulty states which could cause a series of mis-classifications before the model is returned to the true state.

This paper proposes a real-time self-adaptive approach for accurate path loss estimation in underground mines or tunnels based on signal strength measurements from heterogeneous radio communication technologies. The proposed model features simplicity of implementation. The methodology is validated in simulations and verified by measurements taken in real environments. The proposed method leverages accuracy of positioning matching the existing approaches while requiring smaller engineering efforts.

A study investigating the physical properties and use of the SiAl composite Controlled Expansion 7 (CE7) for the packaging of silicon bolometric detectors for millimeter-wave astrophysical applications at cryogenic temperatures is presented. The existing interfaces to such detectors are typically made of either ductile metals or micro-machined silicon. As a composite of Si and Al, we find that CE7 exhibits properties of both in ways that may be advantageous for this application. This exploration of the physical properties of CE7 reveals: (a) superconductivity below a critical transition temperature, T c ∼ 1.2 K; (b) a thermal contraction profile much closer to Si than metal substrates; (c) the relatively low thermal conductivity anticipated for a superconductor, which can be improved by Au-plating; and (d) the feasibility of machining mechanical features with tolerances of ∼25 µm. We further discuss the use of CE7 in the cosmology large angular scale surveyor telescope array, which deployed CE7 in several of its detector focal planes.

Cyclic olefin copolymer (COC) is an amorphous thermoplastic with desirable dielectric and mechanical characteristics for optical applications. In particular, its low refractive index, overall mechanical strength, and absence of strong absorption features make it a promising substrate material for far-infrared applications, which include frequency-selective surfaces, scattering filters, and windows. The dielectric properties of selected COC formulations are surveyed from $ \approx 10 - 700\,\, {{\rm cm}^{ - 1}} $≈10-700cm -1 ($ 1000 - 14\,\, \unicode{x00B5} {\rm m} $1000-14µm), and representative usage as a thin film membrane structure in optical filters is presented.

Potential changes in the contour of the facial profile that accompany tooth movement can be important considerations in developing an orthodontic treatment plan. The objective of this study was to compare the accuracy of three different methods of predicting horizontal soft-tissue changes. Eighty-three nongrowing, orthodontically treated patients comprised the sample. Pretreatment and posttreatment lateral cephalometric hard- and soft-tissue landmark points were digitized. A coordinate system was defined, landmark coordinates were corrected for magnification, and then hard- and soft-tissue, angular, and linear measures were calculated by computer programs developed on a DEC PDP 11/44. The accuracy of prediction of soft-tissue landmark changes was compared for three prediction methods: (1) use of ratios of means of soft-tissue changes to corresponding hard-tissue changes, (2) use of a bivariate regression equation on corresponding hard-tissue landmark changes, and (3) use of stepwise multiple regression with hard-tissue changes and initial hard- and soft-tissue facial characteristics as predictor variables. For predicting changes of four soft-tissue points, multiple regression equations were slightly more accurate than ratio of means predictions. The standard errors of the estimate ranged from 0.7 to 1.1 mm for the multiple regression predictions; these were from 0.2 to 1.4 mm lower than those obtained for ratio of means predictions. The accuracy of the bivariate regression prediction technique fell between that of the other two methods. Examination of the residuals showed that the multiple regression equations consistently underpredicted the most extreme soft-tissue facial changes.

Despite the almost universal agreement on the desirability of bone grafting in the cleft alveolar process, there are little long-term data to support preference for grafting at one time versus another. Investigation was undertaken to compare the clinical success of grafts placed at three distinct developmental stages: the "primary" group consisting of 20 rib grafts placed at less than 1 year of age; the "secondary" group consisting of 19 iliac crest grafts placed when the permanent canine was one fourth to one half formed; and the "delayed" group consisting of 18 iliac crest grafts placed after eruption of the permanent canines. Patients were a minimum of 15 years of age at time of final evaluation. Records were made a minimum of 5 years postsurgery. Biometric data and periapical x-ray films were evaluated. There was a trend for the delayed group to have a decreased incidence of successful bony bridging of the graft site compared with the other treatment groups. The primary group exhibited significantly (P less than 0.001) greater ridge height and increased bone attachment than the secondary and delayed groups. There were significantly fewer teeth lost adjacent to the cleft for the primary group compared with the secondary and delayed groups. Before orthodontic treatment, there were no significant differences among groups in incidence of anterior crossbite; posterior crossbites were seen more frequently in the secondary and delayed graft groups than in the primary group (P less than 0.001). At final evaluation, there were significantly fewer anterior and posterior crossbites in the primary group than in the secondary and delayed groups (P less than 0.002).(ABSTRACT TRUNCATED AT 250 WORDS)