Search

One of the most important ways of improving the speed of complex task solving is employing a multiprocessor computational system. This paper describes the experience of software development for research management and solving educational problems using parallel computing technologies. The authors describe approaches to computation parallelization using a multiprocessor system with shared memory within a task of finding a numerical root of a system of linear equations with a tridiagonal coefficient matrix that appears when solving a boundary problem for a partial differential equation of parabolic type, the heat equation. Additionally, the approaches to parallelization implementation of the tridiagonal matrix method for the heat equation in the two-dimensional case within a numerical root-finding algorithm using the alternating-direction implicit method for a multiprocessor system with shared memory are described. A finite-difference method of variable directions is used to find a numerical root of a heat equation in the two-dimensional case. Sequential and parallel algorithms (two-sided Thomas algorithm and multithread horizontal block Thomas algorithm) that fit an execution on computational systems with shared memory have been used to implement a tridiagonal matrix method. Two parallel computation organization technologies for computational systems with shared memory have been used for computation parallelization: one based on OpenMP technology and one using .NET framework facilities. The parallelization process and load balance serving have been performed by means of the environment in the first case as manual operation of threads parallelization process is allowed in the latter one. As an assessment of the described approach performance, the calculation time for sequential and parallel algorithms is given depending on the task’s size and the number of threads used. Comparison of the considered parallelization algorithms and implementation technologies is performed based on the analysis of the resulting acceleration. This paper shows that total computation time is several times smaller and calculation acceleration is several times bigger when using Thread instead of OpenMP. An application has been developed that allows obtaining a visual result of modelling of process of temperature propagation in the study area using parallel calculation technologies in real time.

The new method for swaging of closed tube ends (cups) based on the sequential process of sheet metal drawing with the subsequent reverse sheet metal drawing on closed tube end is considered. This allows expanding the scope of swaging and using blanks with thickness of 10...15 % less than when drawing by spherical punch. Information about optimal process parameters is presented.

Subject. The state of the stomatognathic system in patients who used removable implant-retained overdentures for a period more than 4 years was analyzed by evaluating the bioelectrical activity of temporal and masticatory muscles. Objectives ― to analyze the influence of long-term usage of removable implant-retained overdentures on the dynamic parameters of the stomatognathic system by evaluating the bioelectrical activity of temporal and masticatory muscles. Methods. Clinical studies of bioelectrical activity and symmetry of the temporal and masticatory muscles using surface electromyography were carried out. Electromyography was carried out in five probes: physiological rest, occlusion, natural swallowing, clench and clench on rolls. Results. It was revealed that in a state of physiological rest and occlusion, the absolute electromyography indicators were elevated. During natural swallowing, the bioelectric activity of the muscles in patients who used removable implant-retained overdentures was characterized as reduced. During clenching and clenching on rolls, the indicators were at the level of the lower limit of the norm. It was also found that the symmetry of the temporal and masticatory muscles was below the normal level. Conclusions. The obtained data is a sign of the modified dynamic state of the stomatognathic system in patients who used removable implant-retained overdentures. In addition to changes of the bioelectric activity of the muscles, a significant violation of the symmetry of the muscles has been revealed.

We developed high energy Q-switched 1064 nm and 946 nm Nd:YAG lasers for ITER Divertor Thomson Scattering (DTS). The 1064 nm Nd:YAG laser is designed to be used as the main DTS probing laser and features output energy of 2 J at 3 ns pulse duration. MOPA configuration with phase conjugation mirror provides divergence of 1.3 times the diffraction limit and high stability of the output beam parameters with repetition rate varied from 1 to 50 Hz. The 946 nm Nd:YAG laser designed for calibration by an auxiliary Thomson spectrum at the neighboring wavelength. The 946 nm laser consists of MO based on cavity dumping, high gain regenerative preamplifier and zero diffraction length ring cavity regenerative power amplifier with image-rotation providing 1 J × 50 Hz at 3 ns pulse duration with divergence of 1.5 the diffraction limit. As far as we know this is the highest value of the output energy obtained at 946 nm with nanosecond pulse duration until now. Both lasers use unified pump modules providing high maintainability of the system. The paper describes laser design approaches, state of the art and outlook of further developments.

We report on a joule-level, 1047 nm Nd:YLF master oscillator power amplifier laser with near-diffraction-limited beam quality. The master oscillator and two-stage two-pass amplifier were all-flashlamp-pumped. The laser produced 2 J output pulses with 3 ns duration at 10 Hz pulse repetition rate. High beam quality was achieved by using a liquid stimulated Brillouin scattering mirror and additional astigmatism compensation.

The paper presents an approach to deal with batch extract-load processes for cloud data lakes. The approach combines multiple data ingestion techniques, provides advanced failover strategies and adopts cloud-native implementation. The suggested approach. The prototype implementation utilizes Amazon Web Services platform and is based on its serverless features. The approach can be implemented also using other cloud platforms like Google Cloud Platform or Microsoft Azure.

Problem-solving lifecycle providing provable semantic interoperability and correct reuse of data, metadata, domain knowledge, methods, and processes on different levels of consideration is proposed. It includes ontological search, data model integration, schema mapping, entity resolution, method and process reuse, hypothesis testing, and data publishing. Problems are solved according to formal domain knowledge specifications over multiple integrated resources. The semantics of every decision may be formally verified.

Nowadays scientific data should be FAIR that are Findable, Accessible, Interoperable and Reusable. Reference implementation of FAIR data management principles proposed recently considers RDF as unifying data model and RDF Mapping Language (RML) as the basic language for data integration. This paper is aimed at development of methods and tools for scalable data integration in the frame of this architecture. A mapping from RML into a high-level data analysis language Pig Latin that runs on Hadoop is considered. The mapping is implemented using model transformation technologies. These allows to execute RML programs in the Hadoop distributed computing environment. According to the experimental evaluation RML implementation developed scales w.r.t. data volume and outperforms related implementations.

We report on a Joule-level flashlamp-pumped 1047 nm Nd:YLF laser with near-diffraction-limited beam quality. The laser was based on master oscillator power amplifier design. The output pulse energy was 2 J with 3 ns pulse duration at 10 Hz pulse repetition rate. High beam quality of the laser output was provided by using an SBS mirror and astigmatism correction in the amplifier.

The analysis of methods and ways to combat forest fires, the results of which were revealed two methods of fire control: direct impact on the edge of the fire (quenching), and methods used at the same time - active methods, and methods related to the deliberate loss Square when creating boom bands and barriers to the spread of fire. These methods are called passive passive or indirect. Active methods used to extinguish fires in the grassroots, and passive - to extinguish the saddle

The design of a forest fire soil-thrower made to prevent and eliminate ground forest fires is presented. A mathematical model of machine movement has been developed, which enables to study the laws of the interaction process of the design with the soil. It is accepted that the machine has two degrees of freedom. The mathematical model has been obtained using the Lagrange equations of the second kind. The design and technological parameters of the forest fire soil-throwing machine, affecting the efficiency of its work, including mass and width of the grip of the ripper casing, mass, radius and frequency of rotation of the milling tool, the number and geometric parameters of the blades are taken into account. Mathematical model enables to determine the effect of these parameters on the characteristics of the movement of ripper casing and milling working body. A mathematical model is needed to synchronize the translational motion of the unit and the rotational motion of the rotor. Formulas have been obtained for the steady motion of the forest fire soil-thrower, that determine the hauling power of tractor and torque that ensures the operation of milling tools.

Glioblastoma is the most common primary adult brain tumour, and despite optimal treatment, the median survival is 12–15 months. Patients with matched recurrent glioblastomas were investigated to try to find actionable mutations. Tumours were profiled using a validated DNA-based gene panel. Copy number variations (CNVs) and single nucleotide variants (SNVs) were examined, and potentially pathogenic variants and clinically actionable mutations were identified. The results revealed that glioblastomas wereIDH-wildtype (IDHWT;n = 38) andIDH-mutant (IDHMUT;n = 3). SNVs inTSC2,MSH6,TP53,CREBBP, andIDH1were variants of unknown significance (VUS) that were predicted to be pathogenic in both subtypes.IDHWTtumours had SNVs that impacted RTK/Ras/PI(3)K, p53, WNT, SHH, NOTCH, Rb, and G-protein pathways. Many tumours hadBRCA1/2(18%) variants, including confirmed somatic mutations in haemangioblastoma.IDHWTrecurrent tumours had fewer pathways impacted (RTK/Ras/PI(3)K, p53, WNT, and G-protein) and CNV gains (BRCA2,GNAS, andEGFR) and losses (TERTandSMARCA4).IDHMUTtumours had SNVs that impacted RTK/Ras/PI(3)K, p53, and WNT pathways. VUS inKLK1was possibly pathogenic inIDHMUT. Recurrent tumours also had fewer pathways (p53, WNT, and G-protein) impacted by genetic alterations. Public datasets (TCGA and GDC) confirmed the clinical significance of findings in both subtypes. Overall in this cohort, potentially actionable variation was most often identified inEGFR,PTEN,BRCA1/2, andATM. This study underlines the need for detailed molecular profiling to identify individual GBM patients who may be eligible for novel treatment approaches. This information is also crucial for patient recruitment to clinical trials.

Background Deciphering the meaning of the human DNA is an outstanding goal which would revolutionize medicine and our way for treating diseases. In recent years, non-coding RNAs have attracted much attention and shown to be functional in part. Yet the importance of these RNAs especially for higher biological functions remains under investigation. Methods In this paper, we analyze RNA-seq data, including non-coding and protein coding RNAs, from lung adenocarcinoma patients, a histologic subtype of non-small-cell lung cancer, with deep learning neural networks and other state-of-the-art classification methods. The purpose of our paper is three-fold. First, we compare the classification performance of different versions of deep belief networks with SVMs, decision trees and random forests. Second, we compare the classification capabilities of protein coding and non-coding RNAs. Third, we study the influence of feature selection on the classification performance. Results As a result, we find that deep belief networks perform at least competitively to other state-of-the-art classifiers. Second, data from non-coding RNAs perform better than coding RNAs across a number of different classification methods. This demonstrates the equivalence of predictive information as captured by non-coding RNAs compared to protein coding RNAs, conventionally used in computational diagnostics tasks. Third, we find that feature selection has in general a negative effect on the classification performance which means that unfiltered data with all features give the best classification results. Conclusions Our study is the first to use ncRNAs beyond miRNAs for the computational classification of cancer and for performing a direct comparison of the classification capabilities of protein coding RNAs and non-coding RNAs.

Efficient and timely fault detection is a significant problem due to the intensifying use of modern technological solutions in machine condition monitoring. This work is carried out as part of a project that is aimed at development of software solutions for a housing and utility condition monitoring system. An experimental setup was designed and assembled for the study of basic housing infrastructure elements operating modes. The setup includes electric pumps, power transformers, ventilation and air conditioning systems (HVAC), heaters and electric boilers. Every element is equipped with various sensors. Sensor readings were gathered, processed and analyzed. This dataset was used to fit statistical and probabilistic models such as linear regression and Hidden Markov model in order to classify regular and faulty operating modes of equipment. Nine classes of equipment malfunction were modeled, these models are intended to be used as a theoretical basis for the design of industrial housing and utility condition monitoring systems.

A new technological process for the synthesis of composite materials based on carbon fibers and the combination of a thermoset and a thermoplastic matrix is described. The mechanical characteristics of a two-matrix composite are presented. The scope of its application is discussed.

The paper presents the research of a methodology of conceptual scheme development to solve problems in subject domains. A semantic approach to domain model specifications building is principal for it. The development process involves formulating a model of requirements to the domain from verbal specifications of domain requirements, developing a domain ontology, transforming it into a conceptual scheme, and reusing domain knowledge specifications in the domain. Relevant data sources are mapped to conceptual schemes of domains in data infrastructures. Requirement specifications are implemented over conceptual schemes for entity resolution and problem-solving in domains using accessible data sources.

According to the Open Science paradigm data sources are to be concentrated within research data infrastructures intended to support the whole cycle of data management and processing. FAIR data management and stewardship principles that had being developed and announced recently state that data within a data infrastructure have to be findable, accessible, interoperable and reusable. Note that data sources can be quite heterogeneous and represented using very different data models. Variety of data models includes traditional relational model and its object-relational extensions, array and graph-based models, semantic models like RDF and OWL, models for semi-structured data like NoSQL, XML, JSON and so on. This particular paper overviews data model unification techniques considered as a formal basis for (meta)data interoperability, integration and reuse within FAIR data infrastructures. These techniques are intended to deal with heterogeneity of data models and their data manipulation languages used to represent data and provide access to data in data sources. General principles of data model unification, languages and formal methods required, stages of data model unification are considered and illustrated by examples. Application of the techniques for data integration within FAIR data infrastructures is discussed.

The work is devoted to the receiver unit calibration of the optoelectronic landing system of an air drone. The optical characteristics of the receiver unit were determined, a calibration stand was developed, a calibration program was created, a calibration measurement series were carried out.

The analysis of the work processes of technical equipment used for forest fire control and preventive activities has revealed that they do not meet the requirements of the process flows. The aim of the work was to increase the efficiency of the work processes of new forest fire-fighting soil-thrower by optimizing the parameters of the milling thrower. The optimization task has been to search for the values of the length and width of the milling thrower blades. These values should provide the highest productivity and smallest supplied power with the average range of soil displacement. Based on the optimization results, the optimal blade length of the milling thrower was revealed in the range of 200-230 mm, and the width of the blade was 150-220 mm. With optimal parameters of the working body, a fire-fighting machine throws soil into the fire zone at a distance of up to 21 m. It consumes no more than 38 kW of power, and its productivity is 0.22 m3/s on the average.

This article dwells upon the investigation of the radiation polarization structure in analysing biological tissues. The systematic investigation of the polarization parameters of lens systems with non-planar media boundaries was performed to increase the accuracy parameters of polarization instruments used to analyse biological tissues. A review of existing solutions, developed mathematical model and experimental setup were given, and series of experiments are described.

During all the period of development of data integration methods and tools the issues of formal semantics definition and verification were arising. Three levels of integration can be distinguished: data model integration, schema matching and integration and data integration proper. This paper is aimed at development of methods and tools for formal semantics definition and verification on the third level – level of data proper. An approach for definition of formal semantics for high-level data integration programs is proposed. The semantics is defined using a transformation into a formal specification language supported by automatic/interactive provers. The semantics is applied for verification of structured data integration workflows. Workflow properties to be verified are presented as expressions of the specification language chosen. After that a semantic specification of the data integration workflow is verified w.r.t. required properties. A practical aim of the work is to define a basis for formal verification of data integration workflows during problem solving in various integration environments.

Amethod and design of an installation for treating iron powders and making preforms for machine parts from them are described. In order to sinter a preform simultaneously with compressing the powder in a closed die pulses of eclectic current are passed through the powder. In accordance with the design of the multifactor experiment, the method is used to fabricate control specimens. Formulas are obtained for computing the relative strain due to upsetting until failure and the specific force of upsetting before failure using the results of measurement of the hardness of articles.

There was defined a formation at the surface of steel billets during the influence of fluctuating at ultrasound frequencies regular nano-structures indicator. The nano-structures form homogeneous micro layer which differs from the original one by smaller fractal dimension. The possibility to form an orientated micro layer at the surface of thermo-processed billets was shown. The micro layer allows to improve the conditions of contacting for the details of the type of matched friction pairs.

Motivation: Data from RNA-seq experiments provide us with many new possibilities to gain insights into biological and disease mechanisms of cellular functioning. However, the reproducibility and robustness of RNA-seq data analysis results is often unclear. This is in part attributed to the two counter acting goals of (i) a cost efficient and (ii) an optimal experimental design leading to a compromise, e.g. in the sequencing depth of experiments. Results: We introduce an R package called samExploreR that allows the subsampling (m out of n bootstraping) of short-reads based on SAM files facilitating the investigation of sequencing depth related questions for the experimental design. Overall, this provides a systematic way for exploring the reproducibility and robustness of general RNA-seq studies. We exemplify the usage of samExploreR by studying the influence of the sequencing depth and the annotation on the identification of differentially expressed genes. Availability and Implementation: samExploreR is available as an R package from Bioconductor. Contact: v@bio-complexity.com Supplementary information: Supplementary data are available at Bioinformatics online.

An approach for applying a combination of the semantically different rule-based languages for interoperable conceptual programming over various rule-based systems (RS) and relying on the logic program transformation technique recommended by the W3C Rule Interchange Format (RIF) is presented. Such approach is coherently combined with the heterogeneous data base integration applying semantic rule mediation. The basic functions of the infrastructure implementing the multi-dialect conceptual specifications by the interoperable RS and mediator programs are defined. The references to the detailed description of the infrastructure application for solving complex combinatorial problem are given. The research results show the usability of the approach and of the infrastructure for declarative, resource independent and re-usable data analysis in various application domains.

Hematopoietic stem cells (HSCs) are rare cells that have the unique ability to self-renew and differentiate into cells of all hematopoietic lineages. The expansion of HSCs has remained an important goal to develop advanced cell therapies for bone marrow transplantation and many blood disorders. Over the last several decades, there have been numerous attempts to expand HSCs in vitro using purified growth factors that are known to regulate HSCs. However, these attempts have been met with limited success for clinical applications. New developments in the HSC expansion field coupled with gene therapy and stem cell transplant should encourage progression in attractive treatment options for many disorders including hematologic conditions, immunodeficiencies, and genetic disorders.

An innovative technique for formation of collaborative consortia of enterprises in virtual organizations (VO) is considered. The technique is based on semantic integration of relevant enterprise information systems (EIS) in the VO specification treated as a subject mediator over the EISs involved. The paper explains how the canonical model for the VO specification is synthesized. Main part of the paper is devoted to the presentation of the process of EIS resource information models semi-automatic mapping into the canonical one assisted by the Heterogeneous Information Model Unifier. It is important to note that the Model Unifier is a universal tool that assists in development of mapping of various kinds of information. The process of mapping includes construction of a compiler from a resource model into the canonical one with the help of metacompilation tools. The mediation technique presented is applied in the astronomical Russian Virtual Observatory (RVO).

Among the most important expected benefits of a global service oriented architecture leveraging web service standards is an increased level of automation in the discovery, composition, verification, monitoring and recovery of services for the realization of complex processes. Most existing works addressing this issue are based on the Ontology Web Language for Services (OWL-S) and founded on description logic. Because the discovery and composition tasks are designed to be fully automatic, the solutions are limited to the realization of rather simple processes. To overcome this deficiency, this paper proposes an approach in which service capability descriptions are based on full first order predicate logic and enable an interactive discovery and composition of services for the realization of complex processes. The proposed approach is well suited when automatic service discovery does not constitute an absolute requirement and the discovery can be done interactively (semi-automatically) with human expert intervention. Such applications are, for instance, often met in e-science. The proposed approach is an extension and adaptation of the compositional information systems development (CISD) method based on the SYNTHESIS language and previously proposed by some of the authors. The resulting method offers a canonical extensible object model with its formal automatic semantic interpretation in the Abstract Machine Notation (AMN) as well as reasoning capabilities applying AMN interactively to the discovery and composition of web services.

The present paper describes an alternative approach to the image forming, which is in a spatial modulation of the gain in the active medium, by the spatial modulation of the density of pump. The problems for the super-Rayleigh resolution are discussed and a possibility of gray scale image generation was shown. The fragmentation of the cross section of the Fabry-Perot cavity helps to suppress smoothening of the image edges during large objects generation. A numerical model of pump modulated lasing pixels (PMLP) was developed, the results of calculation being in good agreement with experiments.

A new principle of the projection display based on a laser with the conjugate resonator is offered where one of the mirrors is a spatial light modulator (SLM). Intra-cavity reading of the image created on the SLM provides the oscillation of an optical signal as well a generated image with high luminosity and optical resolution.

In this paper we describe a simple method of small-size images recognition. The method is based on optical expanded encoding of small-size images by using a random phase-only Fourier filter.© (1999) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

The problem of swift steering of laser beams is important for many applications, (projection, marking, engraving, etc.). An approach of intracavity laser beams spatial control was proposed and developed by several teams [1 -9, 14] . Laser beam movement is provided, with thisapproach, by the positioning of the "ruling diaphragm" across the surface of one of mirrors of conjugate laser cavity. The "ruling diaphragm" can be done as a Spatial Light Modulator (SLM). Such modulators are based on various physical effects and can use various modulating media: electrooptical crystals [2, 10], thin semiconductor layers [1 1], ferro-electrical ceramics [12-14], or liquid crystals [6,8]. The electrically-addressed intracavity Spatial Light Modulators (SLMs) were experimentally shown to provide computer-controlled beam addressing to any pixel insidelaserFieldOfView (FOV) [12, 13,14] In the presented article the following aspects are treated of such laser development: (a) design and properties of iniracavity LC SLMs with random addressing to arbitrary pixel in 2D field, and (b) some peculiarities ofDPSSL with intracavity beam scanning, in terms of SLM operation inside the cavity, and thermal lens influence on the laser FOV

The high-speed laser Fabry-Perot Doppler shift interferometer facility with image recording by means of 8-frame electro-optical camera was elaborated for measuring of the deformation dynamics of the fast moving objects.

Methods of optical design, numerical simulation and experimental testing results are given for the super wide field of vision conjugated optical resonator.

A divertor Thomson scattering (TS) system being developed for ITER has incorporated proven solutions from currently available TS systems. On the other hand any ITER diagnostic has to operate in a hostile environment and very restricted access geometry. Therefore the operation in an environment of intensive stray light, plasma background radiation, the necessity meet the requirement using only a 20 mm gap between divertor cassettes for plasma diagnosis as well as to measure plasma temperatures as low as 1 eV severely constrain the divertor TS diagnostic design. The challenging solutions of this novel diagnostic system which has to ensure its steady performance and also the operability and maintenance are the focus of this report. One of the most demanding parts of the in-vessel diagnostic equipment development is the design assessment using different engineering analyses. The task definition and first results of thermal, e/m and seismic analyses are provided. The process of further improving of the design involves identification of susceptible areas and multiple iterations of the design, as needed. One of the key points for all Thomson scattering diagnostics are the laser capabilities. A high-performance and high-power laser system using a steady-state and high-repetitive mode Nd:YAG laser (2J, 50–100Hz, 3ns) has been developed. The reduced laser pulse duration matched with high-speed low-noise APD detector can be very important under high background light level. For diagnostics such as Thomson scattering and Raman spectroscopy, a high-degree of discrimination against stray light at the laser wavelength is required for successful detection of wavelength-shifted light from the laser-plasma interaction region. For this case of high stray light level, a triple grating polychromator characterized by high rejection and high transmission has been designed and developed. The novel polychromator design minimizes stray light while still maintaining a relatively high transmission.

Figure 1 represents schematically the possible sources of pointing faults. First set of faults is connected with optical distortions and misalignements; these faults exist even when the target does not move. Here are: i - atmospherical distortions caused by turbulence, wind, and sometimes by thermal blooming; L2 distortions and misalignements of optical systems (both receiving, affecting the accuracy of target position determination, and transmitting, thus deteriorating output beam quality) ; E3 -faults that are usually connected with a discrete nature of coordinate photoreceptor; E4 -faults and distortions of the laser optical path: dynamic distortions in amplifiers, scanning systems limited accuracy, misalignements and vibrations of the optical elements, etc. It is worth to note that -2' and L4arenonlinearly dependen on the laser beam power. Second set of faults is connected with temporal delays; they are most dangerous when pointing to the moving targets. Here are: T1 - timeof double light pass to target; T2 - time of obtaining signal from photoreceptors; T3 - time of signal processing; T4 - response time of deflectors. So to achieve the high-stable laser beams pointing means to minimize the values of all listed factors. In some cases physical limitations exist (e.g. one can not overcome the diffraction limit for A2; or the light speed limit for T1 , and it easy to see that for long-distance cases this fundamental limitation becomes the strongest) ; in other cases the break-through lies in the development of non-optical technologies (as, say, in computing technology for T3). Thus our efforts were to find some possible and "optical" ways in order to improve the accuracy of laser beams pointing.

To obtain images, pump radiation is projected onto a thin layer of laser medium through a spatial light modulator. The lasing conditions for such a laser are analyzed, and it is shown that it is possible to generate gray-scale images.

An analysis is made of the temperature dependence of the amplification of pulses of ~10 nsec duration in YAG:Nd3+, assuming that the lifetime of the lower active level is appreciable and allowing for a Boltzman distribution of the sublevel populations of the 4F3/2 and 4I11/2 multiplets. The suitability of this model for the active medium was confirmed experimentally.

An investigation was made of the characteristics of a Q-switched laser utilizing a phosphate glass and producing pulses with a repetition frequency up to 5 sec–1. The pulse duration was 15 nsec when the Q factor of the resonator was switched by an optomechanical modulator with a prism rotating at an equivalent angular velocity of 60,000 rpm.

The results are given of an experimental investigation of diffuse reflectors. It is shown that such reflectors are highly efficient in the case of neodymium-glass lasers. Comparative data are given on the degree of uniformity of the pumping of active elements in specular and diffuse reflectors. The efficiencies of these reflectors are compared. In a system with a diffuse reflector and an immersion layer around the active element the degree of uniformity of the pumping is 0.93 and the efficiency of the reflector is 0.6.

Within scientific and real life problems, classification is a typical case of extremely complex tasks in data-driven scenarios, especially if approached with traditional techniques. Machine Learning supervised and unsupervised paradigms, providing self-adaptive and semi-automatic methods, are able to navigate into large volumes of data characterized by a multi-dimensional parameter space, thus representing an ideal method to disentangle classes of objects in a reliable and efficient way. In Astrophysics, the identification of candidate Globular Clusters through deep, wide-field, single band images, is one of such cases where self-adaptive methods demonstrated a high performance and reliability. Here we experimented some variants of the known Neural Gas model, exploring both supervised and unsupervised paradigms of Machine Learning for the classification of Globular Clusters. Main scope of this work was to verify the possibility to improve the computational efficiency of the methods to solve complex data-driven problems, by exploiting the parallel programming with GPU framework. By using the astrophysical playground, the goal was to scientifically validate such kind of models for further applications extended to other contexts....