2022 - Repozytorium publikacji - Politechnika Gdańska

Pokaż publikacje z roku

Publikacje z roku 2022

Pokaż wszystkie

Neural network training with limited precision and asymmetric exponent
- Marek Blok
- Mariusz Pietrołaj
2022 Pełny tekst Journal of Big Data
Along with an extremely increasing number of mobile devices, sensors and other smart utilities, an unprecedented growth of data can be observed in today’s world. In order to address multiple challenges facing the big data domain, machine learning techniques are often leveraged for data analysis, filtering and classification. Wide usage of artificial intelligence with large amounts of data creates growing demand not only for storage and operational memory, but also computational power. Increasing complexity and variety of neural network architectures are vivid examples of such trends in the modern data-driven industry. In response to this situation, focusing on less demanding operations for inference and training of neural networks became a popular approach among many researchers to overcome resources related issues. This work aims to investigate one of the paths associated with the mentioned efficiency problems and shows the impact of floating-point precision limitation on convolutional neural networks, including experiments on various exponent and mantissa sizes. Additionally, authors explore floating-point numbers utilization and optimization techniques in the scope of neural network training. Based on conducted research a novel method of asymmetric exponent utilization is presented achieving almost identical accuracy of 32-bit floating-point parameters while training a neural network with only 12-bit variables without additional rounding.
Neural Network-Based Sequential Global Sensitivity Analysis Algorithm
- Yen-Chen Liu
- Leifur Leifsson
- Sławomir Kozieł
- Anna Pietrenko-Dąbrowska
2022
Performing global sensitivity analysis (GSA) can be challenging due to the combined effect of the high computational cost, but it is also essential for engineering decision making. To reduce this cost, surrogate modeling such as neural networks (NNs) are used to replace the expensive simulation model in the GSA process, which introduces the additional challenge of finding the minimum number of training data samples required to train the NNs accurately. In this work, a recently proposed NN-based GSA algorithm to accurately quantify the sensitivities is improved. The algorithm iterates over the number of samples required to train the NNs and terminates using an outer-loop sensitivity convergence criteria. The iterative surrogate-based GSA yields converged values for the Sobol’ indices and, at the same time, alleviates the specification of arbitrary accuracy metrics for the NN-based approximation model. In this paper, the algorithm is improved by enhanced NN modeling, which lead to an overall acceleration of the GSA process. The improved algorithm is tested numerically on problems involving an analytical function with three input parameters, and a simulation-based nondestructive evaluation problem with three input parameters.
Neural networks and deep learning
- A. Pastor López-Monroy
- Jesus Garcia Salinas
2022
In this chapter we will provide the general and fundamental background related to Neural Networks and Deep Learning techniques. Specifically, we divide the fundamentals of deep learning in three parts, the first one introduces Deep Feed Forward Networks and the main training algorithms in the context of optimization. The second part covers Convolutional Neural Networks (CNN) and discusses their main advantages and shortcomings for different scenarios and variants of CNNs. Finally, the third part presents Neural Networks for sequence modeling, in particular Recurrent Neural Networks (RNN), Gated Recurrent Units (GRU), Long Short-Term Memory (LSTM) and Attention Mechanisms. The description of the latter models are made in the context of different applications that allows to explain in a better way the details of each particular kind of neural network.
Neuroeconomy and Neuromarketing: The Study of the Consumer Behaviour in the COVID-19 Context
- Juan Jose Blazquez-Resino
- Santiago Gutierrez-Broncano
- Edyta Gołąb-Andrzejak
2022 Pełny tekst Frontiers in Psychology
To address the study of consumer behavior in the post-COVID-19 era, the present Research Topic brings together a set of papers that attempt to study how different factors triggered by the pandemic have achieved a significant effect on consumers' behavioral intentions. These papers examine different subtopics related to food, health products, collaborative economy and, of course, neuroscience. Globally, the objectives of this special issue try to broadly reach the different perspectives in the study of consumer behavior in the context of COVID-19, considering: - A study of the effects of COVID-19 on consumer behavior in the short term. - An analysis of the economic effects of COVID-19. - The social and psychological changes brought about by the current situation of social change. - A study of consumer behavior from the perspective of neuromarketing and/or neuroeconomics. - Forecasting of changes in habits and behaviors caused by COVID-19. - The development of new marketing and distribution systems aimed at meeting new market needs. - New retailing techniques adapted to consumer changes.
Neuroobrazowanie u pacjentów uzależnionych od substancji psychoaktywnych
- Klaudia Malisz
2022 INŻYNIER I FIZYK MEDYCZNY
Celem artykułu jest przedstawienie charakterystycznych zmian w obrębie mózgowia wynikających z zażywania substancji psychoaktywnych, możliwych do zaobserwowania za pomocą badań obrazowych.
Neutral Dissociation of Pyridine Evoked by Irradiation of Ionized Atomic and Molecular Hydrogen Beams
- Tomasz Wąsowicz
2022 Pełny tekst INTERNATIONAL JOURNAL OF MOLECULAR SCIENCES
The interactions of ions with molecules and the determination of their dissociation patterns are challenging endeavors of fundamental importance for theoretical and experimental science. In particular, the investigations on bond-breaking and new bond-forming processes triggered by the ionic impact may shed light on the stellar wind interaction with interstellar media, ionic beam irradiations of the living cells, ion-track nanotechnology, radiation hardness analysis of materials, and focused ion beam etching, deposition, and lithography. Due to its vital role in the natural environment, the pyridine molecule has become the subject of both basic and applied research in recent years. Therefore, dissociation of the gas phase pyridine (C5H5N) into neutral excited atomic and molecular fragments following protons (H+) and dihydrogen cations (H2+) impact has been investigated experimentally in the 5–1000 eV energy range. The collision-induced emission spectroscopy has been exploited to detect luminescence in the wavelength range from 190 to 520 nm at the different kinetic energies of both cations. High-resolution optical fragmentation spectra reveal emission bands due to the CH(A2Δ→X2Πr; B2Σ+→X2Πr; C2Σ+→X2Πr) and CN(B2Σ+→X2Σ+) transitions as well as atomic H and C lines. Their spectral line shapes and qualitative band intensities are examined in detail. The analysis shows that the H2+ irradiation enhances pyridine ring fragmentation and creates various fragments more pronounced than H+ cations. The plausible collisional processes and fragmentation pathways leading to the identified products are discussed and compared with the latest results obtained in cation-induced fragmentation of pyridine.
New 2-[(4-Amino-6-N-substituted-1,3,5-triazin-2-yl)methylthio]-N-(imidazolidin-2-ylidene)-4-chloro-5-methylbenzenesulfonamide Derivatives, Design, Synthesis and Anticancer Evaluation
- Łukasz Tomorowicz
- Beata Żołnowska
- Krzysztof Szafrański
- Jarosław Chojnacki
- Ryszard Konopiński
- Ewa A. Grzybowska
- Jarosław Sławiński
- Anna Kawiak
2022 Pełny tekst INTERNATIONAL JOURNAL OF MOLECULAR SCIENCES
In the search for new compounds with antitumor activity, new potential anticancer agents were designed as molecular hybrids containing the structures of a triazine ring and a sulfonamide fragment. Applying the synthesis in solution, a base of new sulfonamide derivatives 20–162 was obtained by the reaction of the corresponding esters 11–19 with appropriate biguanide hydrochlorides. The structures of the compounds were confirmed by spectroscopy (IR, NMR), mass spectrometry (HRMS or MALDI-TOF/TOF), elemental analysis (C,H,N) and X-ray crystallography. The cytotoxic activity of the obtained compounds toward three tumor cell lines, HCT-116, MCF-7 and HeLa, was examined. The results showed that some of the most active compounds belonged to the R1 = 4-trifluoromethylbenzyl and R1 = 3,5-bis(trifluoromethyl)benzyl series and exhibited IC50 values ranging from 3.6 μM to 11.0 μM. The SAR relationships were described, indicating the key role of the R2 = 4-phenylpiperazin-1-yl substituent for the cytotoxic activity against the HCT-116 and MCF- 7 lines. The studies regarding the mechanism of action of the active compounds included the assessment of the inhibition of MDM2-p53 interactions, cell cycle analysis and apoptosis induction examination. The results indicated that the studied compounds did not inhibit MDM2-p53 interactions but induced G0/G1 and G2/M cell cycle arrest in a p53-independent manner. Furthermore, the active compounds induced apoptosis in cells harboring wild-type and mutant p53. The compound design was conducted step by step and assisted by QSAR models that correlated the activity of the compounds against the HCT-116 cell line with molecular descriptors.
New approach for the synthesis of Ag3PO4-graphene photocatalysts
- Łukasz Lewandowski
- Julia Zwara-Sidorko
- Anna Gołąbiewska
- Tomasz Klimczuk
- Grzegorz Trykowski
- Adriana Zaleska-Medynska
2022 Pełny tekst MATERIALS SCIENCE IN SEMICONDUCTOR PROCESSING
A facile and effective plasma sputtering method for the preparation of a visible light active photocatalyst - rhombic dodecahedral silver phosphate Ag3PO4 covered with nanographene (Ag3PO4-GR) with improved stability has been developed. Proposed method allows for the usage of readily available materials for nanographene sputtering and for easy scaling-up. The stability improvement, confirmed by visible light-induced phenol degradation experiment, could be attributed to the synergistic effect of the silver phosphate particles and graphene material allowing for migration of metallic silver nanoparticles from semiconductor’s surface to graphene body keeping the semiconductors surface “silver free”. Also due to its conductive properties, nanographene may additionally be preventing electron-hole recombination and metallic silver formation.
New Architecture of Solid-State High-Voltage Pulse Generators
- Yahia Achour
- Jacek Starzyński
- Kazimierz Jakubiuk
2022 Pełny tekst ENERGIES
The application of the nanosecond pulsed electric field (nsPEF) for biomedical treatments has gained more interest in recent decades due to the development of pulsed power technologies which provides the ability to control the electric field dose applied during tests. In this context, the proposed paper describes a new architecture of solid-state high-voltage pulse generators (SSHVPG) designed to generate fully customised sequences of quasi-rectangular pulses. The idea is based on the combination of semiconductor switches (IGBT/MOSFET) known for their flexibility and controllability with special magnetic switches to build compact and modular generators. The proposed structure is inspired by the most known pulse generator of Marx, but mixes its two variants for negative and positive polarities. Thus, the polarity of the generated pulses can be freely selected. In addition to that, the use of IGBTs/MOSFET ensures a tunable repetition rate and pulse width. The capacitors are charged via a series of magnetic switches and a flyback DC–DC converter which provides fast and efficient charging and also an adjustable amplitude of the output pulses. The design can be easily simplified giving two other modified structures, based on the same idea, for mono-polar operating (only positive or only negative pulses) with a reduced number of switches. A SPICE simulation of the generator and results of experimental tests carried out on a three stages generator are presented. The obtained results confirm the operating principle and the claimed performances of the new structure.
New insights into modeling two-step nitrification in activated sludge systems – The effects of initial biomass concentrations, comammox and heterotrophic activities
- Mohamad-Javad Mehrani
- Dominika Sobotka
- Przemysław Kowal
- Jianhua Guo
- Jacek Mąkinia
2022 Pełny tekst SCIENCE OF THE TOTAL ENVIRONMENT
In this study, the conventional two-step nitrification model was extended with complete ammonia oxidation (comammox) and heterotrophic denitrification on soluble microbial products. The data for model calibration/validation were collected at four long-term washout experiments when the solid retention time (SRT) and hydraulic retention time (HRT) were progressively reduced from 4 d to 1 d, with mixed liquor suspended solids (MLSS) of approximately 2000 mg/L at the start of each trial. A new calibration protocol was proposed by including a systematic calculation of the initial biomass concentrations and microbial relationships as the calibration targets. Moreover, the impact assessment of initial biomass concentrations (X) and maximum growth rates (µ) for ammonia-oxidizing bacteria (AOB), nitrite-oxidizing bacteria (NOB), comammox Nitrospira, and heterotrophs on the calibration accuracy were investigated using the response surface methodology (RSM). The RSM results revealed the strongest interaction of XAOB and µAOB on the model calibration accuracy. All the examined model efficiency measures confirmed that the extended model was accurately calibrated and validated. The estimated µ values were as follows: µAOB = 0.38±0.005 d-1, µNOB = 0.20±0.01 d-1, µCMX = 0.20±0.01 d-1, µHET = 1.0±0.03 d-1. For comparison, when using the conventional model, µAOB and µNOB increased respectively by 26 and 15% (µAOB = 0.48±0.02 d-1 and µNOB = 0.23±0.005 d-1). This study provides better understanding of the effects of the initial biomass composition and the accompanying processes (comammox and heterotrophic denitrification) on modeling two-step nitrification.
New insights of nanomaterials usage toward superhydrophobic membranes for water desalination via membrane distillation: A review
- Emilia Gontarek-Castro
- Roberto Castro-Muñoz
- Marek Lieder
2022 Pełny tekst CRITICAL REVIEWS IN ENVIRONMENTAL SCIENCE AND TECHNOLOGY
Membrane distillation (MD) is a promising technology for seawater desalination due to the ability to process high-salinity waters and the ability to be driven by low-grade or waste heat. However, practical applications of MD membranes are limited by the low vapor flux and fouling problem. Recently, there is a growing interest in developing novel MD membrane materials with enhanced hydrophobicity to improve the efficiency of desalination performance. Interestingly, the incorporation of nanomaterials for tailoring superhydrophobic properties of MD membranes has attracted enormous attention in MD. Herein, according to the new insights of the available literature data, the current trend for achieving superhydrophobic MD membranes by embedding inorganic nanomaterials is provided. The influence of the inorganic additives on membrane fouling, stability, separation performance, is also discussed. Finally, theoretical principles of MD, the milestones of the evolution of developing superhydrophobic membrane surfaces, and future trends are also given for the new readers in the field.
New nanoadsorbent based on magnetic iron oxide containing 1,4,7,10-tetraazacyclododecane in outer chain (Fe3O4@SiO2-cyclen) for adsorption and removal of selected heavy metal ions Cd2+, Pb2+, Cu2+
- Amanda Kulpa-Koterwa
- Jacek Ryl
- Karolina Górnicka
- Paweł Niedziałkowski
2022 JOURNAL OF MOLECULAR LIQUIDS
Magnetic Fe3O4@SiO2-cyclen nanoparticles were prepared and used as adsorbent for Cd2+, Pb2+ and Cu2+ from aqueous solution removal process controlled with differential pulse anodic stripping voltammetry (DPASV) and hanging mercury drop electrode (HDME). Nanomaterial was synthesised in three-step process co-precipitation of Fe3O4 core, coating with silane and N-(3-(triethoxysilyl)propyl)-1,4,7,10-tetraazacyclododecane-1-carboxamide silane functionalisation. The effectiveness of each step of the synthesis was confirmed using scanning electron microscopy (SEM), high-resolution X-ray photoelectron spectroscopy (XPS), powder X-ray diffraction (pXRD) and fourier-transform infrared spectroscopy (FT-IR) techniques. The Fe3O4@SiO2-cyclen nanoparticles were employed for Cd2+, Pb2+ and Cu2+ ions elimination from individual and mixed solutions by carrying out titration with a suspension of nanocomposites. The binding level for all ions both in the individual solutions and in the mixture was very similar at high levels. For Cd2+ and Cu2+ ions sorption efficiency level was from 83% to 89%, while for Pb2+ ions it was slightly lower at the level over 73%. In all cases, the equilibrium adsorption capacity parameter was over 1 mg/g and reached definitely higher values for individual ions solutions. The research results revealed that Fe3O4@SiO2- cyclen nanoparticles can be a promising adsorbent for magnetic heavy metal ions water treatment agents.
New Peptide Based Fluconazole Conjugates with Expanded Molecular Targets
- Wioletta Brankiewicz
- Joanna Okońska
- Katarzyna Serbakowska
- Jan Lica
- Marek Drab
- Natalia Ptaszyńska
- Anna Łęgowska
- Krzysztof Rolka
- Piotr Szweda
2022 Pełny tekst Pharmaceutics
Infections of Candida spp. etiology are frequently treated with azole drugs. Among azoles, the most widely used in the clinical scenario remains fluconazole (FLC). Promising results in treatment of dangerous, systemic Candida infections demonstrate the advantages of combined therapies carried out with combinations of at least two different antifungal agents. Here, we report five conjugates composed of covalently linked FLC and cell penetrating or antimicrobial peptide: TP10-7-NH2, TP10-NH2, LFcinB(2-11)-NH2, LFcinB[Nle1,11]-NH2, and HLopt2-NH2, with aspects of design, chemical synthesis and their biological activities. Two of these compounds, namely FLCpOH-TP10-NH2 and FLCpOH-TP10-7-NH2, exhibit high activity against reference strains and fluconazole-resistant clinical isolates of C. albicans, including strains overproducing drug transporters. Moreover, both of them demonstrate higher fungicidal effects compared to fluconazole. Analysis performed with fluorescence and scanning electron microscopy as well as flow cytometry indicated the cell membrane as a molecular target of synthesized conjugates. An important advantage of FLCpOH-TP10-NH2 and FLCpOH-TP10-7-NH2 is their low cytotoxicity. The IC90 value for the human cells after 72 h treatment was comparable to the MIC50 value after 24 h treatment for most strains of C. albicans. In reported conjugates, FLC was linked to the peptide by its hydroxyl group. It is worth noting that conjugation of FLC by the nitrogen atom of the triazole ring led to practically inactive compounds. Two compounds produced by us and reported herein appear to be potential candidates for novel antifungal agents.
New Performance Indices for Power System Stabilizers
- Michał Izdebski
- Robert Małkowski
- Piotr Miller
2022 Pełny tekst ENERGIES
The subject of the article is issues related to innovative indices for power system stabilizers (PSSs). These new indices will be able to quickly show which PSS (among many other PSSs) is not working properly and that advanced optimization and simulation methods should be used to improve the PSS settings. The authors note the fact that the acceptance requirements for PSSs are different in various power systems. Moreover, the authors pay attention to the fact that transmission system operators (TSOs) often have different PSS requirements (tests) even though they work in the same large power system. The article reviews the requirements for the PSSs used by TSOs of various power systems. The need to supplement the required tests with new qualitative indices is demonstrated. In the paper, new performance indices are proposed to improve the evaluation of the PSS and to check the desired performance of the stabilizer. These indices are derived from the active power frequency response characteristic with PSS and without PSS (PSS ON and PSS OFF). Additionally, the new PSS indices allow the graphical visualization of the properties of all synchronous generators equipped with the PSS in a predefined area on a single 3D graph. Such visualization can be used to quickly detect weak points of the power system.
New Simple and Robust Method for Determination of Polarity of Deep Eutectic Solvents (DESs) by Means of Contact Angle Measurement
- Łukasz Cichocki
- Dorota Warmińska
- Justyna Łuczak
- Andrzej Przyjazny
- Grzegorz Boczkaj
2022 Pełny tekst MOLECULES
The paper presents a new method for evaluating the polarity and hydrophobicity of deep eutectic solvents (DESs) based on the measurement of the DES contact angle on glass. DESs consisting of benzoic acid derivatives and quaternary ammonium chlorides–tetrabutylammonium chloride (TBAC) and benzyldimethylhexadecylammonium chloride (16-BAC)—in selected molar ratios were chosen for the study. To investigate the DESs polarity, an optical goniometer and an ET(30) solvatochromic scale based on Reichardt’s dye were used. The research demonstrated the high accuracy and precision of the developed procedure. The simplicity of the examination and the availability of basic equipment allow for the implementation of the developed method in routine investigations of DESs.
Nieświadome sieci neuronowe
- Stanisław Barański
2022
Coraz większą popularność zyskuje usługa predykcji za pomocą sieci neuronowych. Model ten zakłada istnienie serwera, który za pomocą wyuczonej sieci neuronowej dokonuje predykcji na danych otrzymanych od klienta. Model ten jest wygodny, ponieważ obie strony mogą skupić się na rozwoju w swojej specjalizacji. Wystawia on jednak na ryzyko utraty prywatności zarówno klienta, wysyła- jącego wrażliwe dane wejściowe, jak i serwer, udostępniający wy- uczony model sieci neuronowej. Niniejszy rozdział opisuje proces dokonywania nieświadomej predykcji za pomocą sieci neuronowych. Nieświadomość pozwala na dokonanie predykcji za pomocą sieci neu- ronowych przy zachowaniu prywatności danych wejściowych klienta oraz modelu serwera. Umożliwia świadczenie usług, które dotych- czas blokowane były przez regulacje prawne lub brak wiarygodności przetwarzających je systemów.
Niobate in silicate and phosphate glasses: Effect of glass basicity on crucible dissolution
- Natalia Wójcik
- Sharafat Ali
- Efstratios Kamitsos
- Doris Möncke
2022 Pełny tekst International Journal of Applied Glass Science
Using niobium crucibles for melting phosphate and silicate glasses of various modifier oxide contents, and therefore varying optical basicity (Λ), was found to result in varying dissolution rates of niobate during melting. Because of their high electronic polarizability, even small concentrations of niobates are detectable in the Raman spectra of glasses. Even less than 1 mol% Nb2O5 can be identified, as independently confirmed by SEM‐EDX analysis. Silica‐rich glasses (~60 % SiO2, Λ~0.6) did not show significant Nb dissolution from the crucible, while higher basicity metasilicate glasses (~ 50% SiO2, Λ~0.65) and pyrophosphate glasses (~30% P2O5, Λ~0.7) did show the typical niobate signature in the Raman spectra at 810‐840 cm‐1, depending on composition. While niobium is well dissolved throughout the pyrophosphate glass, metasilicate glasses showed a much more intense Raman signature of niobate units near the outer surface of the glass. Measurements along the cross section of a fractured metasilicate glass showed a steady decrease of the strength of the niobate signature from the surface toward the bulk of the material. Besides correlation with optical basicity, the tendency of melts to dissolve Nb crucible was discussed in terms of the connectivity or polymerization of the network and the corresponding melt viscosity.
Nitrate and nitrite silver complexes with weakly coordinating nitriles
- Karolina Gutmańska
- Anna Ciborska
- Zbigniew Hnatejko
- Anna Dołęga
2022 Pełny tekst POLYHEDRON
Simple system of weak nitrile ligands and silver nitrite/nitrate allowed the formation and isolation of coordination compounds of diverse nuclearity and dimensionality. The extension of the network does not enhance the luminescent properties of the ligands however it gives rise to the new MLCT bands in the visible region.
NLITED - New Level of Integrated Techniques for Daylighting Education: Preliminary Data on the Use of an E-learning Platform
- Federica Giuliani
- Natalia Sokół
- Niko Gentile,
- Mandana Khanie, Sarey
- Valerio R.M. Lo Verso
- Federica Caffaro
2022
Project NLITED – New Level of Integrated Techniques for Daylighting Education - is an educational project for students and professionals. The project's objective is to create and develop an online eLearning platform with 32 eModules dedicated to daylight knowledge. The project also offers e-learners two summer school training where the theory is put into practice. The platform was launched on January 31, 2022. The paper analyses the participation during the first four months of online activity until May 31, 2022. It discusses which eModules have received the highest participation rate and which have the lowest. These data are compared to the preferences on modules expressed by different panels of experts. The experts expressed their recommendations for specific educational content during workshops conducted in 2021, which led to the definition of the curriculum. Furthermore, participants also fill out an evaluation test on the quality and the usability of the eModule(s) they have taken. This information leads to the amendments of the ePlatform which are in the scope of action for the final year of the NLTED project.
Noise profiling for speech enhancement employing machine learning models
- Krzysztof Kąkol
- Grazina Korvel
- Bożena Kostek
2022 Pełny tekst Journal of the Acoustical Society of America
This paper aims to propose a noise profiling method that can be performed in near real-time based on machine learning (ML). To address challenges related to noise profiling effectively, we start with a critical review of the literature background. Then, we outline the experiment performed consisting of two parts. The first part concerns the noise recognition model built upon several baseline classifiers and noise signal features derived from the Aurora noise dataset. This is to select the best-performing classifier in the context of noise profiling. Therefore, a comparison of all classifier outcomes is shown based on effectiveness metrics. Also, confusion matrices prepared for all tested models are presented. The second part of the experiment consists of selecting the algorithm that scored the best, i.e., Naïve Bayes, resulting in an accuracy of 96.76%, and using it in a noise-type recognition model to demonstrate that it can perform in a stable way. Classification results are derived from the real-life recordings performed in momentary and averaging modes. The key contribution is discussed regarding speech intelligibility improvements in the presence of noise, where identifying the type of noise is crucial. Finally, conclusions deliver the overall findings and future work directions.