The deployment of deep neural network (DNN) models in software applications is increasing rapidly with the exponential growth of artificial intelligence. Currently, such models are deployed manually by developers in the cloud considering several user requirements, while the decision of model selection and user assignment is difficult to take. With the rise of edge computing paradigm, companies tend to deploy applications as close as possible to the user. Considering this system, the problem of DNN model selection and the inference serving becomes harder due to the introduction of communication latency between nodes. We present an automatic method for DNN placement and inference in edge computing; a mathematical formulation to the DNN Model Variant Selection and Placement (MVSP) problem is presented, it considers the inference latency of different model-variants, communication latency between nodes, and utilization cost of edge computing nodes. Furthermore, we propose a general heuristic algorithm to solve the MVSP problem. We provide an analysis of the effects of hardware sharing on inference latency, on an example of GPU edge computing nodes shared between different DNN model-variants. We evaluate our model numerically, and show the potentials of GPU sharing, with decreased average latency by 33% of millisecond-scale per request for low load, and by 21% for high load. We study the tradeoff between latency and cost and show the pareto optimal curves. Finally, we compare the optimal solution with the proposed heuristic and showed that the average latency per request increased by more than 60%. This can be improved using more efficient placement algorithms.

Clinical mistreatment and mismanagement are big issues caused by detection of too many false negative patients. Therefore, lung cancer diagnostic inaccuracy and methods to surpass it in a minimally invasive way is often the subject of research, as it is case of this study. This study focuses on the use of machine learning algorithms as a noninvasive tool to differentiate malignant pleural effusions from benign effusions. It provides performance comparisons between Adaptive neuro-fuzzy inference system (ANFIS), Support vector machine (SVM), RUS Boosted Tree (RUSBoost) and K-Nearest-Neighbor (K-NN) techniques for lung cancer detection. The proposed algorithms were chosen based on the current state of the art in the field of pulmonary diagnostics. The novelty of this work is the application of machine learning models for classification of lung cancer based on expression of tumor markers obtained from serum and pleural fluids. The performance of all models is compared and validated on data samples of 168 patients. Three classification model, SVM, RUSBoost and K-NN performed equally well, whereas underperforming model was ANFIS.

This paper explores the new way of presenting one existing VR application, which was described in our previous work - Virtual Reality Experience of Sarajevo War Heritage. The goal of the application was to introduce more people with the Sarajevo siege and allow them to experience the Tunnel crossing at that time. Before this application, we made two versions, the first one for VR setup and the second for the web. In this paper, we introduce a mobile version with the same content. The challenge was to optimize the content for the mobile experience. The assets were optimized so a wider number of mobile phones with different hardware capabilities can run the application. The advantages and disadvantages of this approach are pointed out, and the limitations of the mobile application are emphasized. The memory usage and frame rate are measured for different Android devices with different operating system versions and hardware capabilities. The results show the optimized application can be run on different Android mobile devices. Nevertheless, for better user experience a higher number of frames per second is needed, which may include reducing the quality of the assets.

Each company aims to remain competitive in the market and provides the services that their clients seek, all in accordance with the cost-effectiveness and fulfillment of customer expectations. In order to do the same, companies are looking for the best practices that help in organizing their work and delivery of their services, as well as maintaining and determining the competitive advantage. In this paper an example of such practice is explained in the ITIL framework. The purpose of the SWOT analysis is to identify strategies that match the resources and capabilities of the company with the needs of the environment in which it competes. The purpose of this analysis is to use the company advantages, to explore its capabilities, to correct weaknesses, and to counteract the threats of the environment. The aim of the paper is to demonstrate the benefits of SWOT analysis with companies with the implemented ITIL framework, which is mainly reflected in simplicity, flexibility, low cost, but with tremendous efficiency, good estimates, and finding negativity and positivism in business.

This paper provides an overview of several standards and gives an example of their implementation in custom visualizations such as dynamic data-driven maps. A specially developed tool MapSpice for data visualization will be presented in this paper. This tool is developed and designed to fulfill accessibility standards and requirements of the web publishing industry. There are descriptions and examples of important features and functionalities in map visualizations that are developed using accessibility best practices. Following these guidelines and recommendations map content is accessible to a larger range of people with different disabilities, including blindness and low vision, photo-sensitivity, speech disabilities, deafness and hearing loss, limited movement, learning disabilities, cognitive limitations, and others.The result of this paper will be an overview of accessibility standards and guidelines and their implementation in custom data-driven maps developed by MapSpice.

Many public figures, companies and associations are planning events in different cities and at the same time have active profiles on social media. The planning process requires processing a large amount of data and different parameters when choosing the best event venue. Social media captures a large number of fan actions per day. This paper describes the process of selecting the most appropriate cities to organize events, aided by data collected from social media. The problem is defined as a combinatorial optimization problem. A modified metaheuristic Bat algorithm was proposed, implemented, and described in detail to solve the problem. Although the original Bat algorithm is designed to solve continuous optimization problems, the implemented bat algorithm is adapted to solve the defined problem. The algorithm is compared to the exhaustive search method for smaller instances, and to the greedy and genetic algorithm for larger instances. The algorithm was tested on benchmark data on cities in 20 European countries, as well as on real data collected from pages on the social network Facebook. Bat algorithm has shown superior results compared to other techniques, both in time and in the quality of the solutions generated.

The rapid development of technology is directly affecting the growth and development of e-commerce shipments, especially in the Business to Customer segment. An increase in e-commerce shipments has a strong impact on the express delivery industry. In these conditions, a very significant challenge is how to organize a postal network. The problem that arises is how many postal centers, and at what locations, should be implemented in a specific geographical area in order to optimize the level of service for the users. Solving this challenge has latterly received increased attention in both industry and academia. The aim of this paper is to firstly provide a concise overview of current approaches in the process of determining the optimal location of postal centers. The second part of the paper will focus on proposing an approach that will rely on machine learning methods for clustering in defined conditions and specific geographical environment using appropriate geographic information tools for spatial data analysis and visualization.

The smart home concept is rapidly becoming a key component in the emerging Internet-of-Things (IoT) society. Home automation systems help customers by improving energy-efficiency, allowing for security monitoring and convenience with simplified control over smart IoT devices. However, it has been determined that the older population has difficulties interacting with complex technical devices. Possible solution to this problem would be involving Interactive Voice Response (IVR) machine, which would enable intelligent smart home control based on the information it gathers from voice-based commands. We explore the concept of a smart home with the integration of voice over IP (VOIP) and IVR technologies, along with smart IoT devices and cloud-based services. The presented smart home concept uses voice-assistance which allows for fluent and intuitive interaction. We have modified existing solutions applied for the English language and accommodated them to work for south Slavic languages. The design and implementation of the prototype for the simple IVR-based smart-home system are explained.

Eco-friendly and rapid microwave processing of a precipitate was used to produce Fe-doped zinc oxide (Zn1-xFexO, x = 0, 0.05, 0.1, 0.15 and 0.20; ZnO:Fe) nanoparticles, which were tested as catalysts toward the oxygen reduction reaction (ORR) and oxygen evolution reaction (OER) in a moderately alkaline solution. The phase composition, crystal structure, morphology, textural properties, surface chemistry, optical properties and band structure were examined to comprehend the influence of Zn2+ partial substitution with Fe3+ on the catalytic activity of ZnO:Fe. Linear sweep voltammetry showed an improved catalytic activity of ZnO:5Fe toward the ORR, compared to pure ZnO, while with increased amounts of the Fe-dopant the activity decreased. The improvement was suggested by a more positive onset potential (0.394 V vs. RHE), current density (0.231 mA cm-2 at 0.150 V vs. RHE), and faster kinetics (Tafel slope, b = 248 mV dec-1), and it may be due to the synergistic effect of (1) a sufficient amount of surface oxygen vacancies, and (2) a certain amount of plate-like particles composed of crystallites with well developed (0001) and (0001[combining macron]) facets. Quite the contrary, the OER study showed that the introduction of Fe3+ ions into the ZnO crystal structure resulted in enhanced catalytic activity of all ZnO:Fe samples, compared to pure ZnO, probably due to the modified binding energy and an optimized band structure. With the maximal current density of 1.066 mA cm-2 at 2.216 V vs. RHE, an onset potential of 1.856 V vs. RHE, and the smallest potential difference between the OER and ORR (ΔE = 1.58 V), ZnO:10Fe may be considered a promising bifunctional catalyst toward the OER/ORR in moderately alkaline solution. This study demonstrates that the electrocatalytic activity of ZnO:Fe strongly depends on the defect chemistry and consequently the band structure. Along with providing fundamental insight into the electrocatalytic activity of ZnO:Fe, the study also indicates an optimal stoichiometry for enhanced bifunctional activity toward the OER/ORR, compared to pure ZnO.

Blockchain technology apparently is a trivial innovation, but this technology has attracted huge investors in a very short period compared to other technologies, and it is still having a lot of potential applications. Smart contracts are making possible execution in an automated and safe way by using blockchain technology. Therefore, smart contracts are applied in this research for the expert system. This paper is about an expert system working with smart contracts and neural networks as the inference machine to decide on the sensors optimal distribution and taking actions when sensor readings are out of range: control lights, activating fire alarms, temperature alarms, etc. for all spaces (parks, schools, hospitals, etc.) in a smart city based on the needs, and likes of the expert system user. This expert system works using a blockchain structure on the EOSIO ecosystem with all data gathered by the sensors being saved in cloud online making internet of things environment and essential data saved in a blockchain node.

We consider the problem of the choice of gauge in nonrelativistic strong-laser-field physics. For this purpose, we use the phase-space path-integral formalism to obtain the momentum-space matrix element of the exact time-evolution operator. With the assumption that the physical transition amplitude corresponds to transitions between eigenstates of the physical energy operator rather than the unperturbed Hamiltonian H0=(−i∂/∂r)2/2+V(r), we prove that the aforementioned momentum-space matrix elements obtained in velocity gauge and length gauge are equal. These results are applied to laser-assisted electron-ion radiative recombination (LAR). The transition amplitude comes out identical in length gauge and velocity gauge, and the expression agrees with the one conventionally obtained in length gauge. In addition to the strong-field approximation (SFA), which is the zeroth-order term of our expansion, we present explicit results for the first-order and the second-order terms, which correspond to LAR preceded by single and double scattering, respectively. Our general conclusion is that in applications to atomic processes in strong-field physics the length-gauge version of the SFA (and its higher-order corrections) should be used. Using the energy operator as the basis-defining Hamiltonian, we have shown that the resulting transition amplitude is gauge invariant and agrees with the form commonly derived in length gauge.

Riverine nutrient loads are among the major causes of eutrophication of the Baltic Sea. This study applied the Soil & Water Assessment Tool (SWAT) in three catchments flowing to the Baltic Sea, namely Vantaanjoki (Finland), Fyrisån (Sweden), and Słupia (Poland), to simulate the effectiveness of nutrient control measures included in the EU’s Water Framework Directive River Basin Management Plans (RBMPs). Moreover, we identified similar, coastal, middle-sized catchments to which conclusions from this study could be applicable. The first modelling scenario based on extrapolation of the existing trends affected the modelled nutrient loads by less than 5%. In the second scenario, measures included in RBMPs showed variable effectiveness, ranging from negligible for Słupia to 28% total P load reduction in Vantaanjoki. Adding spatially targeted measures to RBMPs (third scenario) would considerably improve their effectiveness in all three catchments for both total N and P, suggesting a need to adopt targeting more widely in the Baltic Sea countries.

—The problem of transport optimization is of great importance for the successful operation of distribution companies. To successfully find routes, it is necessary to provide accurate input data on orders, customer location, vehicle fleet, depots, and delivery restrictions. Most of the input data can be provided through the order creation process or the use of various online services. One of the most important inputs is an estimate of the unloading time of the goods for each customer. The number of customers that the vehicle serves during the day directly depends on the time of unloading. This estimate depends on the number of items, weight and volume of orders, but also on the specifics of customers, such as the proximity of parking or crowds at the unloading location. Customers repeat over time, and unloading time can be calculated from GPS data history. The paper describes the innovative application of machine learning techniques and delivery history obtained through a GPS vehicle tracking system for a more accurate estimate of unloading time. The application of techniques gave quality results and significantly improved the accuracy of unloading time data by 83.27% compared to previously used methods. The proposed method has been implemented for some of the largest distribution companies in Bosnia and Herzegovina.

Background & Aims YAP (Yap1) and TAZ (Wwtr1) are transcriptional co-activators and downstream effectors of the Hippo pathway, which play crucial roles in organ size control and cancer pathogenesis. Genetic deletion of YAP/TAZ has shown their critical importance for embryonic development of the heart, vasculature, and gastrointestinal mesenchyme. The aim of this study was to determine the functional role of YAP/TAZ in adult smooth muscle cells in vivo. Methods Because YAP and TAZ are mutually redundant, we used YAP/TAZ double-floxed mice crossed with mice that express tamoxifen-inducible CreERT2 recombinase driven by the smooth muscle–specific myosin heavy chain promoter. Results Double-knockout of YAP/TAZ in adult smooth muscle causes lethality within 2 weeks, mainly owing to colonic pseudo-obstruction, characterized by severe distension and fecal impaction. RNA sequencing in colon and urinary bladder showed that smooth muscle markers and muscarinic receptors were down-regulated in the YAP/TAZ knockout. The same transcripts also correlated with YAP/TAZ in the human colon. Myograph experiments showed reduced contractility to depolarization by potassium chloride and a nearly abolished muscarinic contraction and spontaneous activity in colon rings of YAP/TAZ knockout. Conclusions YAP and TAZ in smooth muscle are guardians of colonic contractility and control expression of contractile proteins and muscarinic receptors. The knockout model has features of human chronic intestinal pseudo-obstruction and may be useful for studying this disease.

To develop a consensus on diagnosis and treatment of acromioclavicular joint instability. A consensus process following the modified Delphi technique was conducted. Panel members were selected among the European Shoulder Associates of ESSKA. Five rounds were performed between October 2018 and November 2019. The first round consisted of gathering questions which were then divided into blocks referring to imaging, classifications, surgical approach for acute and chronic cases, conservative treatment. Subsequent rounds consisted of condensation by means of an online questionnaire. Consensus was achieved when ≥ 66.7% of the participants agreed on one answer. Descriptive statistic was used to summarize the data. A consensus was reached on the following topics. Imaging: a true anteroposterior or a bilateral Zanca view are sufficient for diagnosis. 93% of the panel agreed on clinical override testing during body cross test to identify horizontal instability. The Rockwood classification, as modified by the ISAKOS statement, was deemed valid. The separation line between acute and chronic cases was set at 3 weeks. The panel agreed on arthroscopically assisted anatomic reconstruction using a suspensory device (86.2%), with no need of a biological augmentation (82.8%) in acute injuries, whereas biological reconstruction of coracoclavicular and acromioclavicular ligaments with tendon graft was suggested in chronic cases. Conservative approach and postoperative care were found similar A consensus was found on the main topics of controversy in the management of acromioclavicular joint dislocation. Each step of the diagnostic treatment algorithm was fully investigated and clarified. Level V.