Heterostructures of the materials system GaN/AlGaN/GaN were grown by molecular beam epitaxy on 6H-SiC substrates and high electron mobility transistors (HEMTs) were fabricated. For devices with large gate periphery an air bridge technology was developed for the drain contacts of the finger structure. The devices showed DC drain currents of more than 1 A/mm and values of the transconductance between 120 and 140 mS/mm. A power added efficiency of 41 % was measured on devices with a gate length of 1 µm at 2 GHz and 45 V drain bias. Power values of 8 W/mm were obtained. Devices with submicron gates exhibited power values of 6.1 W/mm (7 GHz) and 3.16 W/mm (25 GHz) respectively. The rf dispersion of the drain current is very low, although the devices were not passivated. / Heterostrukturen im Materialsystem GaN/AlGaN/GaN wurden mittels Molekularstrahlepitaxie auf 6H-SiC-Substraten gewachsen und High-Electron-Mobility-Transistoren (HEMTs) daraus hergestellt. Für Bauelemente mit großer Gateperipherie wurde eine Air-Bridge-Technik entwickelt, um die Drainkontakte der Fingerstruktur zu verbinden. Die Bauelemente zeigten Drainströme von mehr als 1 A/mm und Steilheiten zwischen 120 und 140 mS/mm. An Transistoren mit Gatelängen von 1 µm konnten Leistungswirkungsgrade (Power Added Efficiency) von 41 % (bei 2 GHz und 45 V Drain-Source-Spannung) sowie eine Leistung von 8 W/mm erzielt werden. Bauelemente mit Gatelängen im Submikrometerbereich zeigten Leistungswerte von 6,1 W/mm (7 GHz) bzw. 3,16 W/mm (25 GHz). Die Drainstromdispersion ist sehr gering, obwohl die Bauelemente nicht passiviert wurden.

In this thesis, a custom quadrupole mass filter setup was established to independently investigate the impact of the most prominent ion species that are present during ion-assisted deposition. The setup was applied to the low temperature epitaxial growth of GaN thin films on 6H-SiC substrates. Atomic nitrogen ions at higher ion kinetic energies were for the first time independently identified to be the predominant cause of deteriorating crystalline qualities during growth. Precise control of the ion beam parameters yielded the capability to vary the average GaN phase content from almost purely wurtzite to the meta-stable zinc blende GaN phase. Even in case of comparably high crystalline quality, the atomic and molecular nitrogen ions were independently determined to yield distinct thin film topographies throughout the entire observed evolution of the thin film formation.:Bibliographical Description 1 Introduction 1.1 Epitaxial Thin Film Growth 1.2 Ion-Beam Assisted Deposition 1.2.1 Influence of Energetic Particles 1.2.2 Ion-atom Arrival Ratio 1.3 Gallium Nitride 2 Methods 2.1 Setup of the Deposition System 2.1.1 Knudsen Effusion Cell 2.1.2 Reflection High-Energy Electron Diffraction 2.1.3 Auger Electron Spectroscopy 2.1.4 Ion Sources 2.2 Quadrupole Mass Filter System 2.2.1 Components 2.2.2 Working Principle of a Quadrupole Mass Filter 2.2.3 Alternative Mass Filters 2.3 X-ray Diffraction and Reflectivity 2.4 Atomic Force Microscopy 2.5 Transmission Electron Microscopy 3 Results and Discussions 3.1 Characterization of the Quadrupole Mass Filter System 3.1.1 Mass Filter Performance and Resolution 3.1.2 Ion Beam Characteristics 3.1.3 Space Charge Considerations 3.1.4 Conclusions 3.2 Influence of the I/A Ratio and Ion Kinetic Energy 3.2.1 Determination of the GaN Phase Composition 3.2.2 Film Topography and Growth Mode 3.2.3 Crystal Structure and Orientation 3.2.4 Microstructure at the Interface 3.2.5 Conclusions 3.3 Impact of the Ion Species on Growth Instabilities 3.3.1 Growth Rates and Thin Film Topography 3.3.2 Crystal Structure 3.3.3 Growth Mode and RHEED pattern evolution 3.3.4 Conclusions 4 Summary and Conclusions Bibliography Complete Publication List of the Author Acknowledgments Declaration of Authorship / In dieser Arbeit wurde ein maßgefertigter Quadrupol-Massenfilteraufbau etabliert, um die Auswirkungen der prominentesten Ionenspezies, die während der ionengestützten Abscheidung vorhanden sind, unabhängig voneinander zu untersuchen. Der Aufbau wurde für das epitaktische Niedertemperatur-Wachstum von GaN-Dünnschichten auf 6H-SiC-Substraten angewendet. Atomare Stickstoffionen bei höheren kinetischen Ionenenergien wurden zum ersten Mal in der Abwesenheit anderer Spezies als die dominierende Ursache für die Verschlechterung der kristallinen Qualität während des Wachstums identifiziert. Eine präzise Kontrolle der Ionenstrahlparameter ergab die Fähigkeit, den durchschnittlichen GaN-Phasengehalt von der fast reinen Wurtzit- bis zur metastabilen Zinkblende-GaN-Phase zu variieren. Selbst bei vergleichbar hoher kristalliner Qualität weisen die mit atomaren und molekularen Stickstoffionen hergestellten Schichten unabhängig voneinander verschiedene Topographien auf, die sich während der gesamten beobachteten Entwicklung der Dünnschichtbildung deutlich abzeichneten.:Bibliographical Description 1 Introduction 1.1 Epitaxial Thin Film Growth 1.2 Ion-Beam Assisted Deposition 1.2.1 Influence of Energetic Particles 1.2.2 Ion-atom Arrival Ratio 1.3 Gallium Nitride 2 Methods 2.1 Setup of the Deposition System 2.1.1 Knudsen Effusion Cell 2.1.2 Reflection High-Energy Electron Diffraction 2.1.3 Auger Electron Spectroscopy 2.1.4 Ion Sources 2.2 Quadrupole Mass Filter System 2.2.1 Components 2.2.2 Working Principle of a Quadrupole Mass Filter 2.2.3 Alternative Mass Filters 2.3 X-ray Diffraction and Reflectivity 2.4 Atomic Force Microscopy 2.5 Transmission Electron Microscopy 3 Results and Discussions 3.1 Characterization of the Quadrupole Mass Filter System 3.1.1 Mass Filter Performance and Resolution 3.1.2 Ion Beam Characteristics 3.1.3 Space Charge Considerations 3.1.4 Conclusions 3.2 Influence of the I/A Ratio and Ion Kinetic Energy 3.2.1 Determination of the GaN Phase Composition 3.2.2 Film Topography and Growth Mode 3.2.3 Crystal Structure and Orientation 3.2.4 Microstructure at the Interface 3.2.5 Conclusions 3.3 Impact of the Ion Species on Growth Instabilities 3.3.1 Growth Rates and Thin Film Topography 3.3.2 Crystal Structure 3.3.3 Growth Mode and RHEED pattern evolution 3.3.4 Conclusions 4 Summary and Conclusions Bibliography Complete Publication List of the Author Acknowledgments Declaration of Authorship

Light Detection And Ranging (LiDAR) is a hot topic today primarily because of its vast importance within autonomous vehicles. LiDAR sensors are capable of capturing and identifying objects in the 3D environment. However, a drawback of LiDAR is that they perform poorly under adverse weather conditions. Noise present in LiDAR scans can be divided into random and pseudo-random noise. Random noise can be modeled and mitigated by statistical means. The same approach works on pseudo-random noise, but it is less effective. For this, Deep Neural Nets (DNN) are better suited. The main goal of this thesis is to investigate how snow can be detected in LiDAR point clouds and filtered out. The dataset used is Winter Adverse DrivingdataSet (WADS). Supervised filtering contains a comparison between statistical filtering and segmentation-based neural networks and is evaluated on recall, precision, and F1. The supervised approach is expanded by investigating an ensemble approach. The supervised result indicates that neural networks have an advantage over statistical filters, and the best result was obtained from the 3D convolution network with an F1 score of 94.58%. Our ensemble approaches improved the F1 score but did not lead to more snow being removed. We determine that an ensemble approach is a sub-optimal way of increasing the prediction performance and holds the drawback of being more complex. We also investigate an unsupervised approach. The unsupervised networks are evaluated on their ability to find noisy data and correct it. Correcting the LiDAR data means predicting new values for detected noise instead of just removing it. Correctness of such predictions is evaluated manually but with the assistance of metrics like PSNR and SSIM. None of the unsupervised networks produced an acceptable result. The reason behind this negative result is investigated and presented in our conclusion, along with a model that suffers none of the flaws pointed out.

Malware is becoming more intelligent as static methods for blocking communication with Command and Control (C&C) server are becoming obsolete. Domain Generation Algorithms (DGAs) are a common evasion technique that generates pseudo-random domain names to communicate with C&C servers in a difficult way to detect using handcrafted methods. Trying to detect DGAs by looking at the domain name is a broad and efficient approach to detect malware-infected hosts. This gives us the possibility of detecting a wider assortment of malware compared to other techniques, even without knowledge of the malware’s existence. Our study compared the effectiveness of three different machine learning classifiers: Convolutional Neural Network (CNN), Generative Adversarial Network (GAN) and Random Forest (RF) when recognizing patterns and identifying these pseudo-random domains. The result indicates that CNN differed significantly from GAN and RF. It achieved 97.46% accuracy in the final evaluation, while RF achieved 93.89% and GAN achieved 60.39%. In the future, network traffic (efficiency) could be a key component to examine, as productivity may be harmed if the networkis over burdened by domain identification using machine learning algorithms.

No description available.

Denna artikel utforskar artificiell intelligens (AI) som en kreativ partner inom digitalbildproduktion och dess potential att förändra och komplettera traditionella designmetoder.Artikeln undersöker AI:s roll inom designprocessen, främst inriktat mot AI-baserade systemsom kan syntetisera visuellt material, diskuterar dess styrkor, svagheter och begränsningarsamt reflekterar över hur AI påverkar kreativitet, effektivisering och idégenerering. Som ettkomplement till artikeln har en fotobok skapats för att på ett visuellt engagerande sättpresentera resultaten av en AI:s försök till tolkning av textbeskrivningar, med hjälp avAI-verktyget Midjourney och GPT-4. Vidare granskas hur AI-drivna designprocesser kansamverka med traditionella metoder och på så sätt belysa hur AI kan integreras i designerskreativa verktygslåda och därmed utnyttjas för att skapa intressanta och unika resultat.Artikeln bidrar till förståelsen av AI:s växande roll inom designområdet och erbjuder insikter ihur designers kan använda AI som en kreativ partner.

Motivated by successes in the image generation domain, this thesis presents a novel Hybrid VQ-VAE (H-VQ-VAE) approach for generating realistic synthetic time series data with categorical features. The primary motivation behind this work is to address the limitations of existing generative models in accurately capturing the underlying structure and patterns of time series data, especially when dealing with categorical features.  Our proposed H-VQ-VAE model builds upon the foundation of the VQ-VAE architecture and consists of two separate VQ-VAEs: the whole VQ-VAE and the sliding VQ-VAE. Both models share a ResNet-based architecture with conv1d layers to effectively capture the temporal structure within the time series data. The whole VQ-VAE focuses on entire sequences of data to learn relationships between categorical and numerical features, while the sliding VQ-VAE exclusively processes numerical features using a sliding window approach. We conducted experiments on multiple datasets to evaluate the performance of our H-VQ-VAE model in comparison with the original VQ-VAE and TimeGAN models. Our evaluation used a train-on-real and test-on-synthetic approach, focusing on metrics such as Mean Absolute Error (MAE) and Explained Variance (EV). The H-VQ-VAE model achieved a 25-50% better MAE for numerical features compared to the VQ-VAE and outperformed TimeGAN by 45-75% on the complex dataset indicating its effectiveness in capturing the underlying structure and patterns of the time series data. In conclusion, the H-VQ-VAE model offers a promising approach for generating realistic synthetic time series data with categorical features, with potential applications in various fields where accurate data generation is crucial.

Data augmentation is a technique that acquires more training data by augmenting available samples, where the training data is used to fit model parameters. Data augmentation is utilized due to a shortage of training data in certain domains and to reduce overfitting. Augmenting a training dataset for image classification with a Generative Adversarial Network (GAN) has been shown to increase classification accuracy. This report investigates if transfer learning within a GAN can further increase classification accuracy when utilizing the augmented training dataset. The method section describes a specific GAN architecture for the experiments that includes a label condition. When using transfer learning within the specific GAN architecture, a statistical analysis shows a statistically significant increase in classification accuracy for a classification problem with the EMNIST dataset, which consists of images of handwritten alphanumeric characters. In the discussion section, the authors analyze the results and motivates other use cases for the proposed GAN architecture. / Datautökning är en metod som skapar mer träningsdata genom att utöka befintlig träningsdata, där träningsdatan används för att anpassa modellers parametrar. Datautökning används på grund av en brist på träningsdata inom vissa områden samt för att minska overfitting. Att utöka ett träningsdataset för att genomföra bildklassificering med ett generativt adversarialt nätverk (GAN) har visats kunna öka precisionen av klassificering av bilder. Denna rapport undersöker om transferlärande inom en GAN kan vidare öka klassificeringsprecisionen när ett utökat träningsdataset används. Metoden beskriver en specific GANarkitektur som innehåller ett etikettvillkor. När transferlärande används inom den utvalda GAN-arkitekturen visar en statistisk analys en statistiskt säkerställd ökning av klassificeringsprecisionen för ett klassificeringsproblem med EMNIST datasetet, som innehåller bilder på handskrivna bokstäver och siffror. I diskussionen diskuteras orsakerna bakom resultaten och fler användningsområden nämns.

The thesis proposes an unsupervised representation learning method to predict 3D human pose from a 2D skeleton via a VAEGAN (Variational Autoencoder Generative Adversarial Network) hybrid network. The method learns to lift poses from 2D to 3D using selfsupervision and adversarial learning techniques. The method does not use images, heatmaps, 3D pose annotations, paired/unpaired 2Dto3D skeletons, 3D priors, synthetic 2D skeletons, multiview or temporal information in any shape or form. The 2D skeleton input is taken by a VAE that encodes it in a latent space and then decodes that latent representation to a 3D pose. The 3D pose is then reprojected to 2D for a constrained, selfsupervised optimization using the input 2D pose. Parallelly, the 3D pose is also randomly rotated and reprojected to 2D to generate a ’novel’ 2D view for unconstrained adversarial optimization using a discriminator network. The combination of the optimizations of the original and the novel 2D views of the predicted 3D pose results in a ’realistic’ 3D pose generation. The thesis shows that the encoding and decoding process of the VAE addresses the major challenge of erroneous and incomplete skeletons from 2D detection networks as inputs and that the variance of the VAE can be altered to get various plausible 3D poses for a given 2D input. Additionally, the latent representation could be used for crossmodal training and many downstream applications. The results on Human3.6M datasets outperform previous unsupervised approaches with less model complexity while addressing more hurdles in scaling the task to the real world. / Uppsatsen föreslår en oövervakad metod för representationslärande för att förutsäga en 3Dpose från ett 2D skelett med hjälp av ett VAE GAN (Variationellt Autoenkodande Generativt Adversariellt Nätverk) hybrid neuralt nätverk. Metoden lär sig att utvidga poser från 2D till 3D genom att använda självövervakning och adversariella inlärningstekniker. Metoden använder sig vare sig av bilder, värmekartor, 3D poseannotationer, parade/oparade 2D till 3D skelett, a priori information i 3D, syntetiska 2Dskelett, flera vyer, eller tidsinformation. 2Dskelettindata tas från ett VAE som kodar det i en latent rymd och sedan avkodar den latenta representationen till en 3Dpose. 3D posen är sedan återprojicerad till 2D för att genomgå begränsad, självövervakad optimering med hjälp av den tvådimensionella posen. Parallellt roteras dessutom 3Dposen slumpmässigt och återprojiceras till 2D för att generera en ny 2D vy för obegränsad adversariell optimering med hjälp av ett diskriminatornätverk. Kombinationen av optimeringarna av den ursprungliga och den nya 2Dvyn av den förutsagda 3Dposen resulterar i en realistisk 3Dposegenerering. Resultaten i uppsatsen visar att kodningsoch avkodningsprocessen av VAE adresserar utmaningen med felaktiga och ofullständiga skelett från 2D detekteringsnätverk som indata och att variansen av VAE kan modifieras för att få flera troliga 3D poser för givna 2D indata. Dessutom kan den latenta representationen användas för crossmodal träning och flera nedströmsapplikationer. Resultaten på datamängder från Human3.6M är bättre än tidigare oövervakade metoder med mindre modellkomplexitet samtidigt som de adresserar flera hinder för att skala upp uppgiften till verkliga tillämpningar.

No description available.